| *** ivar-lazzaro has quit IRC | 00:00 | |
| *** ivar-lazzaro has joined #openstack-neutron | 00:01 | |
| *** dannywilson has quit IRC | 00:01 | |
| *** ivar-laz_ has joined #openstack-neutron | 00:05 | |
| *** ivar-lazzaro has quit IRC | 00:05 | |
| *** rangansw has quit IRC | 00:05 | |
| *** zhenq has joined #openstack-neutron | 00:07 | |
| *** baoli has joined #openstack-neutron | 00:07 | |
| *** xgerman has quit IRC | 00:08 | |
| *** salv-orlando has quit IRC | 00:10 | |
| *** dave-mccowan has joined #openstack-neutron | 00:13 | |
| *** tgohad has quit IRC | 00:15 | |
| *** leenheer has quit IRC | 00:16 | |
| openstackgerrit | Arvind Somya proposed openstack/neutron: Model and migration for cisco providernet extension driver https://review.openstack.org/171326 | 00:20 |
|---|---|---|
| *** rook__ has joined #openstack-neutron | 00:24 | |
| openstackgerrit | Kevin Benton proposed openstack/neutron: Add simple ARP spoofing protection https://review.openstack.org/171003 | 00:24 |
| *** jaypipes has quit IRC | 00:27 | |
| *** rook__ has quit IRC | 00:29 | |
| *** ronis_ has quit IRC | 00:30 | |
| *** dane_leblanc has quit IRC | 00:41 | |
| *** claudiub has quit IRC | 00:43 | |
| *** Zemeio has quit IRC | 00:44 | |
| *** emagana has quit IRC | 00:46 | |
| prometheanfire | ijw: we didn't get that far in checking | 00:49 |
| prometheanfire | ijw: I think the next step for us is to see if we can manually change the host bridge/tap interface's mtu, if that works we know the neutron-agent/server peices are working, and then apply a test patch | 00:50 |
| *** Longgeek has joined #openstack-neutron | 00:50 | |
| prometheanfire | to the client | 00:50 |
| Sam-I-Am | prometheanfire: moo | 00:51 |
| prometheanfire | Sam-I-Am: mooooo | 00:51 |
| prometheanfire | Sam-I-Am: does that plan sound good? | 00:51 |
| *** pavel_bondar has quit IRC | 00:51 | |
| prometheanfire | Sam-I-Am: btw, I'll need the IP, since it's my work laptop that's still ssh'd in | 00:52 |
| Sam-I-Am | prometheanfire: i did some more testing and don't actually see the mtu being set on device creation | 00:52 |
| *** pavel_bondar has joined #openstack-neutron | 00:52 | |
| Sam-I-Am | however, in the agents container, they're set somehow... and not on compute. | 00:52 |
| *** Marga_ has quit IRC | 00:53 | |
| prometheanfire | Sam-I-Am: it's because the agent on the nova host/compute/whatever doesn | 00:53 |
| prometheanfire | doesn't run through ml2, but I could be talking out my ass | 00:53 |
| Sam-I-Am | prometheanfire: 23.253.65.75 | 00:53 |
| Sam-I-Am | it should be using ml2 | 00:53 |
| *** nlahouti has quit IRC | 00:53 | |
| Sam-I-Am | the lb agents all use the same config | 00:53 |
| prometheanfire | it should, but at least the linuxbridge section isn't being called thoguh it I think | 00:53 |
| prometheanfire | the LB ini file has many sections | 00:54 |
| prometheanfire | main one is where you set the mtu | 00:54 |
| Sam-I-Am | i sure hope its reading it all... | 00:54 |
| prometheanfire | but I think the nova server only cares about the linuxbridge section | 00:54 |
| prometheanfire | at least that's what that patch implies | 00:54 |
| prometheanfire | to me | 00:54 |
| *** oda-g has quit IRC | 00:55 | |
| *** oda-g has joined #openstack-neutron | 00:56 | |
| *** Sukhdev has quit IRC | 00:58 | |
| Sam-I-Am | prometheanfire: this needs more digging | 00:58 |
| Sam-I-Am | or just some answers in here might be easier :) | 00:58 |
| *** amuller has quit IRC | 00:59 | |
| prometheanfire | Sam-I-Am: I still think it testing by manually settng the mtu of the nova-comupte's client bridge/tap to verify that the neutron-server/client stuff is good would be benificial | 00:59 |
| prometheanfire | it would allow us to focus solely on the nova-compute side | 00:59 |
| *** zzzeek has quit IRC | 01:00 | |
| Sam-I-Am | also seems like bridges assume mtu of the smallest port on them, and wont let you change it from 1500 unless those devices are higher | 01:01 |
| prometheanfire | I think it's smallest by default, but can be set seperately | 01:03 |
| prometheanfire | the order I did the host was tap -> bridge | 01:03 |
| prometheanfire | so I don't know for sure | 01:03 |
| Sam-I-Am | it should be tap -> bridge | 01:03 |
| *** shwetaap has quit IRC | 01:04 | |
| prometheanfire | I mean the order I changed mtu manually | 01:04 |
| Sam-I-Am | right | 01:05 |
| *** Longgeek has quit IRC | 01:07 | |
| *** _cjones_ has quit IRC | 01:08 | |
| *** Longgeek has joined #openstack-neutron | 01:08 | |
| prometheanfire | Sam-I-Am: you are right about the ordering, just tested | 01:09 |
| *** banix has joined #openstack-neutron | 01:13 | |
| dmz | is it possible to have a per-network setting for dhcp-agent to send differnet "search" settings for dns settings; currently resolv.conf gets set to search openstack, i want to have it insert a custom search domain per subnet | 01:15 |
| openstackgerrit | Saksham Varma proposed openstack/neutron: N1Kv: Make neutron-VSM sync interval configurable https://review.openstack.org/171451 | 01:17 |
| openstackgerrit | Saksham Varma proposed openstack/neutron: N1Kv: Make neutron-VSM sync interval configurable https://review.openstack.org/171451 | 01:21 |
| *** shwetaap has joined #openstack-neutron | 01:23 | |
| *** banix has quit IRC | 01:23 | |
| *** yamamoto has joined #openstack-neutron | 01:24 | |
| *** shwetaap1 has joined #openstack-neutron | 01:24 | |
| *** ivar-laz_ has quit IRC | 01:25 | |
| *** rook__ has joined #openstack-neutron | 01:25 | |
| *** signed8bit has joined #openstack-neutron | 01:28 | |
| *** Kennan has joined #openstack-neutron | 01:28 | |
| Kennan | hi neutron-cores, could I ask one question ? | 01:28 |
| *** shwetaap has quit IRC | 01:28 | |
| *** ijw has quit IRC | 01:28 | |
| Kennan | I found if I used latest devstack, when the step to install neutron side | 01:29 |
| *** signed8bit has quit IRC | 01:29 | |
| Kennan | the machine would lost connection | 01:29 |
| Kennan | Does neutron have new changes to devstack ? like loadbalance and etc. | 01:29 |
| Kennan | Seems neutron side issue | 01:29 |
| *** rook__ has quit IRC | 01:29 | |
| *** sbfox has joined #openstack-neutron | 01:31 | |
| *** luqas has joined #openstack-neutron | 01:31 | |
| *** leenheer has joined #openstack-neutron | 01:32 | |
| *** s3wong has quit IRC | 01:33 | |
| *** popw has joined #openstack-neutron | 01:33 | |
| *** stanzgy has joined #openstack-neutron | 01:34 | |
| *** luqas has quit IRC | 01:36 | |
| *** dboik has joined #openstack-neutron | 01:37 | |
| *** haleyb_ has joined #openstack-neutron | 01:37 | |
| *** alex_xu has joined #openstack-neutron | 01:37 | |
| *** tiswanso has joined #openstack-neutron | 01:38 | |
| *** tiswanso has quit IRC | 01:38 | |
| *** tiswanso has joined #openstack-neutron | 01:38 | |
| *** achanda has quit IRC | 01:38 | |
| *** melwitt has quit IRC | 01:40 | |
| *** erkules has quit IRC | 01:42 | |
| *** tiswanso has quit IRC | 01:42 | |
| *** tiswanso has joined #openstack-neutron | 01:43 | |
| *** vivek-eb_ has quit IRC | 01:44 | |
| gus | cores: can someone approve https://review.openstack.org/#/c/160230/ already? I'm getting sick of patching it into all my changes :/ | 01:47 |
| gus | (forces tox to run tests under python2.7, instead of whatever version of python `tox` runs with. The Debian python-tox package has moved to python3 recently) | 01:48 |
| gus | to be more precise: s/tests/pep8, etc/ | 01:48 |
| *** dims_ has quit IRC | 01:51 | |
| *** sbfox has quit IRC | 01:51 | |
| *** baoli has quit IRC | 01:52 | |
| *** baoli has joined #openstack-neutron | 01:52 | |
| *** tomoe_ has joined #openstack-neutron | 01:56 | |
| *** jckasper has joined #openstack-neutron | 01:57 | |
| *** tflynn has joined #openstack-neutron | 01:58 | |
| *** salv-orlando has joined #openstack-neutron | 01:59 | |
| *** tiswanso_ has joined #openstack-neutron | 01:59 | |
| *** jdaggett has quit IRC | 02:00 | |
| *** haleyb_ has quit IRC | 02:01 | |
| *** tiswanso has quit IRC | 02:01 | |
| *** comstud has quit IRC | 02:01 | |
| *** lbragstad has quit IRC | 02:01 | |
| *** Trozz_ has joined #openstack-neutron | 02:01 | |
| *** boris-42 has quit IRC | 02:02 | |
| *** rstarr has quit IRC | 02:02 | |
| *** adam_g has quit IRC | 02:02 | |
| *** gus has quit IRC | 02:02 | |
| *** zacksh has quit IRC | 02:02 | |
| *** Trozz has quit IRC | 02:02 | |
| *** jroll has quit IRC | 02:02 | |
| *** gus has joined #openstack-neutron | 02:03 | |
| *** zacksh has joined #openstack-neutron | 02:03 | |
| *** salv-orlando has quit IRC | 02:04 | |
| *** jdaggett has joined #openstack-neutron | 02:04 | |
| *** boris-42 has joined #openstack-neutron | 02:05 | |
| *** adam_g has joined #openstack-neutron | 02:07 | |
| *** comstud has joined #openstack-neutron | 02:07 | |
| *** jroll has joined #openstack-neutron | 02:07 | |
| *** adam_g has quit IRC | 02:07 | |
| *** adam_g has joined #openstack-neutron | 02:07 | |
| *** jroll has quit IRC | 02:08 | |
| *** jroll has joined #openstack-neutron | 02:08 | |
| *** lbragstad has joined #openstack-neutron | 02:11 | |
| *** julim has quit IRC | 02:12 | |
| *** yamahata has quit IRC | 02:16 | |
| *** vivek-ebay has joined #openstack-neutron | 02:17 | |
| *** jckasper has quit IRC | 02:17 | |
| *** chandankumar has joined #openstack-neutron | 02:22 | |
| *** harlowja is now known as harlowja_away | 02:22 | |
| *** baoli has quit IRC | 02:24 | |
| *** tflynn has quit IRC | 02:24 | |
| *** rook__ has joined #openstack-neutron | 02:26 | |
| *** sbfox has joined #openstack-neutron | 02:26 | |
| *** chandankumar has quit IRC | 02:27 | |
| *** baoli has joined #openstack-neutron | 02:28 | |
| *** banix has joined #openstack-neutron | 02:29 | |
| *** tflynn has joined #openstack-neutron | 02:29 | |
| *** rook__ has quit IRC | 02:30 | |
| *** tomoe_ has quit IRC | 02:36 | |
| *** tflynn has quit IRC | 02:37 | |
| *** tflynn has joined #openstack-neutron | 02:38 | |
| *** jckasper has joined #openstack-neutron | 02:38 | |
| *** achanda has joined #openstack-neutron | 02:39 | |
| *** suresh12 has quit IRC | 02:44 | |
| *** tflynn has quit IRC | 02:44 | |
| *** flwang has quit IRC | 02:44 | |
| *** flwang has joined #openstack-neutron | 02:44 | |
| *** banix has quit IRC | 02:45 | |
| *** jckasper has quit IRC | 02:48 | |
| *** jckasper has joined #openstack-neutron | 02:49 | |
| *** dims has joined #openstack-neutron | 02:51 | |
| *** tomoe_ has joined #openstack-neutron | 02:52 | |
| *** rstarr has joined #openstack-neutron | 02:53 | |
| *** baoli has quit IRC | 02:56 | |
| *** dims has quit IRC | 02:57 | |
| *** jckasper has quit IRC | 02:58 | |
| openstackgerrit | YAMAMOTO Takashi proposed openstack/neutron: test_ovs_neutron_agent: Remove useless ofport=10 arguments https://review.openstack.org/158208 | 02:58 |
| *** tflynn has joined #openstack-neutron | 03:00 | |
| *** sbfox has quit IRC | 03:04 | |
| *** leenheer has quit IRC | 03:05 | |
| *** vivek-ebay has quit IRC | 03:06 | |
| *** julim has joined #openstack-neutron | 03:08 | |
| *** tflynn has quit IRC | 03:11 | |
| openstackgerrit | YAMAMOTO Takashi proposed openstack/neutron: test_l2population: Reduce code duplication in migration tests https://review.openstack.org/158184 | 03:13 |
| openstackgerrit | YAMAMOTO Takashi proposed openstack/neutron: test_l2population: Simplify migration tests https://review.openstack.org/158182 | 03:13 |
| *** xgerman has joined #openstack-neutron | 03:16 | |
| *** boris-42 has quit IRC | 03:18 | |
| *** xgerman has quit IRC | 03:18 | |
| *** zhenq has quit IRC | 03:19 | |
| *** tiswanso_ has quit IRC | 03:19 | |
| *** vivek-ebay has joined #openstack-neutron | 03:19 | |
| *** alex_xu has quit IRC | 03:20 | |
| *** alex_xu has joined #openstack-neutron | 03:21 | |
| *** rook__ has joined #openstack-neutron | 03:26 | |
| *** yuanying has joined #openstack-neutron | 03:28 | |
| *** salv-orlando has joined #openstack-neutron | 03:28 | |
| *** vivek-ebay has quit IRC | 03:30 | |
| *** vivek-ebay has joined #openstack-neutron | 03:31 | |
| *** rook__ has quit IRC | 03:31 | |
| *** luqas has joined #openstack-neutron | 03:33 | |
| *** luqas has quit IRC | 03:38 | |
| *** tflynn has joined #openstack-neutron | 03:38 | |
| *** xgerman has joined #openstack-neutron | 03:43 | |
| *** achanda has quit IRC | 03:43 | |
| *** shwetaap1 has quit IRC | 03:44 | |
| *** coolsvap|afk is now known as coolsvap | 03:48 | |
| *** xgerman has quit IRC | 03:54 | |
| *** leenheer has joined #openstack-neutron | 03:55 | |
| *** pradipta has joined #openstack-neutron | 03:55 | |
| openstackgerrit | RaoFei proposed openstack/neutron: Moving VLAN Transparency support from core to extension https://review.openstack.org/169569 | 03:56 |
| *** tomoe__ has joined #openstack-neutron | 03:58 | |
| *** tomoe_ has quit IRC | 03:58 | |
| *** shwetaap has joined #openstack-neutron | 03:58 | |
| *** vivek-ebay has quit IRC | 04:00 | |
| *** pal has joined #openstack-neutron | 04:01 | |
| openstackgerrit | Assaf Muller proposed openstack/neutron: Update port bindings for master router https://review.openstack.org/141114 | 04:02 |
| *** sthillma has joined #openstack-neutron | 04:04 | |
| *** YorikSar has joined #openstack-neutron | 04:04 | |
| openstackgerrit | pritesh proposed openstack/neutron: Moving VLAN Transparency support from core to extension https://review.openstack.org/169569 | 04:05 |
| *** pritesh has joined #openstack-neutron | 04:05 | |
| *** psd has joined #openstack-neutron | 04:06 | |
| *** tomoe__ has quit IRC | 04:07 | |
| *** rwsu has quit IRC | 04:07 | |
| openstackgerrit | Assaf Muller proposed openstack/neutron: Update port bindings for master router https://review.openstack.org/141114 | 04:07 |
| *** salv-orlando has quit IRC | 04:08 | |
| *** tomoe_ has joined #openstack-neutron | 04:09 | |
| *** erkules has joined #openstack-neutron | 04:11 | |
| *** dboik has quit IRC | 04:13 | |
| *** suresh12 has joined #openstack-neutron | 04:16 | |
| *** _cjones_ has joined #openstack-neutron | 04:19 | |
| *** _cjones_ has quit IRC | 04:19 | |
| *** _cjones_ has joined #openstack-neutron | 04:20 | |
| *** aranjan has joined #openstack-neutron | 04:21 | |
| *** aswadr has joined #openstack-neutron | 04:21 | |
| *** seizadi has joined #openstack-neutron | 04:22 | |
| *** zhhuabj_ has joined #openstack-neutron | 04:22 | |
| *** anish_ has joined #openstack-neutron | 04:23 | |
| *** shwetaap has quit IRC | 04:24 | |
| *** shwetaap has joined #openstack-neutron | 04:25 | |
| *** zhhuabj has quit IRC | 04:26 | |
| *** tomoe_ has quit IRC | 04:27 | |
| *** rook__ has joined #openstack-neutron | 04:27 | |
| *** leenheer has quit IRC | 04:28 | |
| *** _cjones_ has quit IRC | 04:28 | |
| *** tomoe_ has joined #openstack-neutron | 04:29 | |
| *** rook__ has quit IRC | 04:32 | |
| *** tflynn has quit IRC | 04:32 | |
| *** psd has quit IRC | 04:34 | |
| *** sbfox has joined #openstack-neutron | 04:35 | |
| *** chandankumar has joined #openstack-neutron | 04:40 | |
| *** irenab has joined #openstack-neutron | 04:40 | |
| openstackgerrit | Merged openstack/neutron: Fix intermittent UT failures in test_utils https://review.openstack.org/171433 | 04:40 |
| *** coolsvap is now known as coolsvap|afk | 04:41 | |
| *** YorikSar has quit IRC | 04:43 | |
| *** achanda has joined #openstack-neutron | 04:44 | |
| *** anshul has joined #openstack-neutron | 04:44 | |
| *** Marga_ has joined #openstack-neutron | 04:45 | |
| *** Marga_ has quit IRC | 04:45 | |
| *** Marga_ has joined #openstack-neutron | 04:45 | |
| *** Marga_ has quit IRC | 04:45 | |
| *** Marga_ has joined #openstack-neutron | 04:46 | |
| *** aranjan has quit IRC | 04:46 | |
| *** aranjan has joined #openstack-neutron | 04:47 | |
| *** xgerman has joined #openstack-neutron | 04:48 | |
| *** annegentle has joined #openstack-neutron | 04:49 | |
| *** achanda has quit IRC | 04:49 | |
| *** neeti has joined #openstack-neutron | 04:50 | |
| *** neeti has quit IRC | 04:50 | |
| *** neeti has joined #openstack-neutron | 04:50 | |
| *** trinaths has joined #openstack-neutron | 04:52 | |
| *** psd has joined #openstack-neutron | 04:53 | |
| *** coolsvap|afk is now known as coolsvap | 04:53 | |
| *** annegentle has quit IRC | 04:54 | |
| *** TrevorV|Home has joined #openstack-neutron | 04:57 | |
| *** yuanying has quit IRC | 04:58 | |
| *** tomoe_ has quit IRC | 05:00 | |
| *** aranjan has quit IRC | 05:01 | |
| *** dave-mccowan has quit IRC | 05:02 | |
| *** xgerman has quit IRC | 05:03 | |
| *** TrevorV|Home has quit IRC | 05:05 | |
| *** lazy_prince has joined #openstack-neutron | 05:05 | |
| *** salv-orlando has joined #openstack-neutron | 05:11 | |
| *** julim has quit IRC | 05:15 | |
| *** tsg has joined #openstack-neutron | 05:20 | |
| openstackgerrit | Gal Sagie proposed openstack/neutron: OVS_LIB API addition - change bridge controller connection-mode https://review.openstack.org/167913 | 05:21 |
| openstackgerrit | Hirofumi Ichihara proposed openstack/neutron-specs: Add enable_new_agents to agent conf in network node https://review.openstack.org/170774 | 05:22 |
| *** yatin has joined #openstack-neutron | 05:24 | |
| *** armax has quit IRC | 05:26 | |
| *** achanda has joined #openstack-neutron | 05:26 | |
| *** suresh12 has quit IRC | 05:28 | |
| *** rook__ has joined #openstack-neutron | 05:28 | |
| *** changbl has joined #openstack-neutron | 05:31 | |
| *** rook__ has quit IRC | 05:33 | |
| *** tomoe_ has joined #openstack-neutron | 05:34 | |
| *** nlahouti has joined #openstack-neutron | 05:36 | |
| openstackgerrit | Merged openstack/neutron: Remove tests from HA routers test framework https://review.openstack.org/171040 | 05:37 |
| *** abregman has joined #openstack-neutron | 05:40 | |
| *** bobmel has joined #openstack-neutron | 05:40 | |
| *** nlahouti has quit IRC | 05:40 | |
| *** nlahouti has joined #openstack-neutron | 05:40 | |
| *** dims has joined #openstack-neutron | 05:41 | |
| *** dims has quit IRC | 05:47 | |
| *** salv-orlando has quit IRC | 05:49 | |
| *** emagana has joined #openstack-neutron | 05:50 | |
| *** bradjones has quit IRC | 05:52 | |
| *** YorikSar has joined #openstack-neutron | 05:52 | |
| *** ronis_ has joined #openstack-neutron | 05:54 | |
| *** ankit_ag has joined #openstack-neutron | 05:54 | |
| *** seizadi has quit IRC | 05:55 | |
| *** nlahouti has quit IRC | 05:56 | |
| *** nplanel_ has quit IRC | 05:57 | |
| *** bradjones has joined #openstack-neutron | 05:57 | |
| *** bradjones has quit IRC | 05:57 | |
| *** bradjones has joined #openstack-neutron | 05:57 | |
| *** Tiancheng has joined #openstack-neutron | 05:59 | |
| *** vthapar has joined #openstack-neutron | 06:03 | |
| *** aepifanov has joined #openstack-neutron | 06:04 | |
| *** nlahouti has joined #openstack-neutron | 06:04 | |
| openstackgerrit | Edgar Magana proposed openstack/neutron: Add missing config parameters in neutron.conf https://review.openstack.org/171059 | 06:06 |
| *** nlahouti has quit IRC | 06:06 | |
| *** nlahouti has joined #openstack-neutron | 06:07 | |
| *** apuimedo has joined #openstack-neutron | 06:07 | |
| *** Murali has joined #openstack-neutron | 06:08 | |
| openstackgerrit | Merged openstack/python-neutronclient: Support fwaasrouterinsertion extension https://review.openstack.org/166776 | 06:09 |
| openstackgerrit | Adolfo Duarte proposed openstack/neutron: Fix DVR FloatingIP fail because of weakref leakage https://review.openstack.org/159230 | 06:18 |
| *** kiran_ has joined #openstack-neutron | 06:22 | |
| *** _kiran_ has joined #openstack-neutron | 06:22 | |
| *** kiran_ is now known as kiran-r | 06:26 | |
| *** rook_ has joined #openstack-neutron | 06:29 | |
| *** anshul has quit IRC | 06:31 | |
| *** anshul has joined #openstack-neutron | 06:31 | |
| *** gildub has quit IRC | 06:31 | |
| *** rook_ has quit IRC | 06:33 | |
| *** kawa2014 has joined #openstack-neutron | 06:35 | |
| *** psd has quit IRC | 06:37 | |
| *** psd has joined #openstack-neutron | 06:37 | |
| *** sthillma has quit IRC | 06:38 | |
| *** jprovazn has joined #openstack-neutron | 06:39 | |
| *** suresh12 has joined #openstack-neutron | 06:39 | |
| *** abehl has joined #openstack-neutron | 06:41 | |
| *** anshul has quit IRC | 06:41 | |
| *** emagana has quit IRC | 06:42 | |
| *** emagana has joined #openstack-neutron | 06:42 | |
| *** suresh12 has quit IRC | 06:43 | |
| *** yfauser has joined #openstack-neutron | 06:46 | |
| *** gsagie_ has joined #openstack-neutron | 06:46 | |
| *** emagana has quit IRC | 06:47 | |
| *** nyechiel__ has joined #openstack-neutron | 06:48 | |
| *** yfauser has left #openstack-neutron | 06:48 | |
| *** doude has joined #openstack-neutron | 06:49 | |
| *** tsg has quit IRC | 06:52 | |
| *** saju_m has joined #openstack-neutron | 06:52 | |
| *** moshele has joined #openstack-neutron | 06:56 | |
| *** anshul has joined #openstack-neutron | 06:56 | |
| *** abehl has quit IRC | 06:58 | |
| *** erkules has quit IRC | 06:58 | |
| *** erkules has joined #openstack-neutron | 06:58 | |
| *** banix has joined #openstack-neutron | 07:01 | |
| *** nlahouti has quit IRC | 07:02 | |
| *** subscope has quit IRC | 07:04 | |
| openstackgerrit | pritesh proposed openstack/neutron: Moving VLAN Transparency support from core to extension https://review.openstack.org/169569 | 07:07 |
| *** numan has joined #openstack-neutron | 07:08 | |
| *** shwetaap has quit IRC | 07:10 | |
| *** sbfox has quit IRC | 07:12 | |
| *** kickinz1|afk is now known as kickinz1 | 07:12 | |
| *** salv-orlando has joined #openstack-neutron | 07:14 | |
| *** achanda has quit IRC | 07:15 | |
| *** achanda has joined #openstack-neutron | 07:16 | |
| *** subscope has joined #openstack-neutron | 07:17 | |
| *** akamyshnikova has quit IRC | 07:19 | |
| *** pask81 has joined #openstack-neutron | 07:19 | |
| *** matrohon has joined #openstack-neutron | 07:23 | |
| *** numan has quit IRC | 07:23 | |
| *** luqas has joined #openstack-neutron | 07:23 | |
| *** afazekas has joined #openstack-neutron | 07:23 | |
| *** pask81 has quit IRC | 07:25 | |
| *** luqas has quit IRC | 07:26 | |
| *** pask81 has joined #openstack-neutron | 07:29 | |
| *** anilvenkata has joined #openstack-neutron | 07:29 | |
| *** rook_ has joined #openstack-neutron | 07:29 | |
| *** gsagie_ has quit IRC | 07:29 | |
| *** anshul has quit IRC | 07:29 | |
| *** jlibosva has joined #openstack-neutron | 07:31 | |
| *** pritesh has quit IRC | 07:31 | |
| *** akamyshnikova has joined #openstack-neutron | 07:31 | |
| *** luqas has joined #openstack-neutron | 07:32 | |
| *** tflynn has joined #openstack-neutron | 07:33 | |
| *** Marga_ has quit IRC | 07:33 | |
| *** rook_ has quit IRC | 07:34 | |
| *** tflynn has quit IRC | 07:37 | |
| *** anshul has joined #openstack-neutron | 07:38 | |
| *** yog__ has joined #openstack-neutron | 07:39 | |
| *** achanda has quit IRC | 07:40 | |
| openstackgerrit | Merged openstack/neutron-vpnaas: VPNaaS: Remove check for bash usage https://review.openstack.org/171278 | 07:42 |
| *** jpena has joined #openstack-neutron | 07:49 | |
| *** amotoki has joined #openstack-neutron | 07:50 | |
| *** salv-orl_ has joined #openstack-neutron | 07:50 | |
| *** ygbo has joined #openstack-neutron | 07:51 | |
| *** safchain has joined #openstack-neutron | 07:51 | |
| *** SridharG1 has joined #openstack-neutron | 07:53 | |
| *** salv-orlando has quit IRC | 07:55 | |
| *** Murali has quit IRC | 07:57 | |
| *** huwei_ has joined #openstack-neutron | 08:00 | |
| *** huwei__ has joined #openstack-neutron | 08:01 | |
| *** jistr has joined #openstack-neutron | 08:02 | |
| *** banix has quit IRC | 08:06 | |
| *** atuvenie has joined #openstack-neutron | 08:06 | |
| *** aepifanov has quit IRC | 08:11 | |
| *** YorikSar has quit IRC | 08:13 | |
| *** bradjones has quit IRC | 08:14 | |
| *** YorikSar has joined #openstack-neutron | 08:14 | |
| *** ivar-lazzaro has joined #openstack-neutron | 08:20 | |
| *** bradjones has joined #openstack-neutron | 08:20 | |
| *** bradjones has quit IRC | 08:21 | |
| *** bradjones has joined #openstack-neutron | 08:21 | |
| *** mikal has quit IRC | 08:22 | |
| *** mikal has joined #openstack-neutron | 08:23 | |
| *** ihrachyshka has joined #openstack-neutron | 08:25 | |
| *** saju_m has quit IRC | 08:27 | |
| *** rook_ has joined #openstack-neutron | 08:30 | |
| openstackgerrit | YAMAMOTO Takashi proposed openstack/neutron: l2pop UT: Reduce code duplication in migration tests https://review.openstack.org/158184 | 08:34 |
| openstackgerrit | YAMAMOTO Takashi proposed openstack/neutron: l2pop UT: Simplify migration tests https://review.openstack.org/158182 | 08:34 |
| openstackgerrit | YAMAMOTO Takashi proposed openstack/neutron: l2pop UT: Expire cached db objects before reusing a session https://review.openstack.org/171520 | 08:34 |
| *** ivar-lazzaro has quit IRC | 08:34 | |
| *** rook_ has quit IRC | 08:34 | |
| *** oomichi has joined #openstack-neutron | 08:37 | |
| *** salv-orl_ has quit IRC | 08:47 | |
| *** atuvenie has quit IRC | 08:49 | |
| *** anshul has quit IRC | 08:56 | |
| *** ronis_ has quit IRC | 08:56 | |
| *** nyechiel__ is now known as nyechiel | 08:59 | |
| openstackgerrit | Mathieu Rohon proposed openstack/neutron: ML2: Change port status only when it's bound to the host https://review.openstack.org/163178 | 09:00 |
| *** luqas has quit IRC | 09:01 | |
| *** Murali has joined #openstack-neutron | 09:05 | |
| *** tomoe_ has quit IRC | 09:07 | |
| *** ociuhandu has quit IRC | 09:11 | |
| *** tomoe_ has joined #openstack-neutron | 09:14 | |
| *** anshul has joined #openstack-neutron | 09:16 | |
| *** moshele has quit IRC | 09:18 | |
| *** dims has joined #openstack-neutron | 09:19 | |
| *** hichihara has quit IRC | 09:20 | |
| *** dims has quit IRC | 09:24 | |
| *** pradipta has quit IRC | 09:30 | |
| *** anish_ has quit IRC | 09:31 | |
| *** rook_ has joined #openstack-neutron | 09:31 | |
| *** luqas has joined #openstack-neutron | 09:31 | |
| *** johnchalekson has joined #openstack-neutron | 09:35 | |
| *** rook_ has quit IRC | 09:36 | |
| *** kawa2014 has quit IRC | 09:37 | |
| *** johnchalekson has quit IRC | 09:38 | |
| *** johnchalekson has joined #openstack-neutron | 09:40 | |
| *** johnchalekson has quit IRC | 09:40 | |
| *** johnchalekson has joined #openstack-neutron | 09:41 | |
| *** johnchalekson has quit IRC | 09:41 | |
| *** johnchalekson has joined #openstack-neutron | 09:43 | |
| *** ZZelle has joined #openstack-neutron | 09:43 | |
| *** anish_ has joined #openstack-neutron | 09:43 | |
| *** johnchalekson has joined #openstack-neutron | 09:44 | |
| *** johnchalekson has quit IRC | 09:44 | |
| *** johnchalekson has joined #openstack-neutron | 09:45 | |
| *** johnchalekson has joined #openstack-neutron | 09:45 | |
| openstackgerrit | Kevin Benton proposed openstack/neutron: Add simple ARP spoofing protection https://review.openstack.org/171003 | 09:46 |
| *** johnchalekson has joined #openstack-neutron | 09:46 | |
| *** johnchalekson has joined #openstack-neutron | 09:47 | |
| openstackgerrit | Pranav Salunke proposed openstack/neutron: Add non-isolated network metadata server route https://review.openstack.org/159580 | 09:50 |
| *** rook_ has joined #openstack-neutron | 09:52 | |
| *** kawa2014 has joined #openstack-neutron | 09:52 | |
| *** kawa2014 has quit IRC | 09:55 | |
| *** kawa2014 has joined #openstack-neutron | 09:55 | |
| *** nmagnezi has joined #openstack-neutron | 09:56 | |
| *** nmagnezi has quit IRC | 09:57 | |
| openstackgerrit | Kevin Benton proposed openstack/neutron: Add simple ARP spoofing protection https://review.openstack.org/171003 | 09:57 |
| *** yamamoto has quit IRC | 09:58 | |
| *** Murali has quit IRC | 09:58 | |
| openstackgerrit | Kevin Benton proposed openstack/neutron: Add simple ARP spoofing protection https://review.openstack.org/171003 | 09:59 |
| *** ronis has joined #openstack-neutron | 10:00 | |
| *** ociuhandu has joined #openstack-neutron | 10:00 | |
| *** kickinz1 is now known as kickinz1|afk | 10:03 | |
| *** kickinz1|afk is now known as kickinz1 | 10:04 | |
| *** rm_work has quit IRC | 10:04 | |
| *** raildo|away has quit IRC | 10:06 | |
| *** rm_work has joined #openstack-neutron | 10:06 | |
| *** atuvenie has joined #openstack-neutron | 10:08 | |
| *** dims has joined #openstack-neutron | 10:08 | |
| *** salv-orlando has joined #openstack-neutron | 10:08 | |
| *** raildo|away has joined #openstack-neutron | 10:09 | |
| *** yog__ has quit IRC | 10:10 | |
| *** tomoe_ has quit IRC | 10:11 | |
| *** zhhuabj_ has quit IRC | 10:13 | |
| *** zhhuabj has joined #openstack-neutron | 10:14 | |
| *** jprovazn has quit IRC | 10:23 | |
| *** rook_ has quit IRC | 10:24 | |
| *** jprovazn has joined #openstack-neutron | 10:24 | |
| *** pc_m has joined #openstack-neutron | 10:25 | |
| *** ferest has joined #openstack-neutron | 10:25 | |
| *** yog__ has joined #openstack-neutron | 10:26 | |
| *** ebagdasa_ has quit IRC | 10:27 | |
| *** wznoinsk has quit IRC | 10:29 | |
| *** wznoinsk has joined #openstack-neutron | 10:31 | |
| *** ferest has quit IRC | 10:31 | |
| *** pdb has quit IRC | 10:32 | |
| *** pdb has joined #openstack-neutron | 10:32 | |
| *** moshele has joined #openstack-neutron | 10:37 | |
| *** numan has joined #openstack-neutron | 10:41 | |
| *** claudiub has joined #openstack-neutron | 10:51 | |
| *** stanzgy has quit IRC | 10:52 | |
| *** f13o has joined #openstack-neutron | 10:52 | |
| *** ebagdasa_ has joined #openstack-neutron | 10:53 | |
| *** yamamoto has joined #openstack-neutron | 10:53 | |
| *** annegentle has joined #openstack-neutron | 10:53 | |
| *** ebagdasa_1 has joined #openstack-neutron | 10:54 | |
| *** johnchalekson has quit IRC | 10:56 | |
| *** johnchalekson has joined #openstack-neutron | 10:57 | |
| *** ebagdasa_ has quit IRC | 10:58 | |
| *** annegentle has quit IRC | 10:58 | |
| *** johnchalekson has joined #openstack-neutron | 10:59 | |
| *** ebagdasa_ has joined #openstack-neutron | 11:01 | |
| *** ebagdasa_1 has quit IRC | 11:01 | |
| openstackgerrit | Sridhar Gaddam proposed openstack/neutron: Spawn RADVD only in the master HA router https://review.openstack.org/171570 | 11:02 |
| *** ebagdasa_ has quit IRC | 11:04 | |
| *** mwagner_lap has quit IRC | 11:05 | |
| *** ebagdasa_ has joined #openstack-neutron | 11:06 | |
| *** claudiub has quit IRC | 11:06 | |
| openstackgerrit | Grzegorz Grasza (xek) proposed openstack/python-neutronclient: Update neutronclient shell to use shared arguments from Session https://review.openstack.org/170370 | 11:08 |
| *** jpena has quit IRC | 11:09 | |
| pc_m | Has anyone started working on fixing FWaaS for neutron import changes? https://bugs.launchpad.net/neutron/+bug/1441501 | 11:09 |
| openstack | Launchpad bug 1441501 in neutron "gate-neutron-fwaas-python27 fails with ImportError: cannot import name test_db_plugin" [Critical,New] | 11:09 |
| *** johnchalekson has joined #openstack-neutron | 11:09 | |
| *** johnchalekson has quit IRC | 11:09 | |
| *** ebagdasa_ has quit IRC | 11:11 | |
| *** ebagdasa_ has joined #openstack-neutron | 11:11 | |
| *** yog__ has quit IRC | 11:12 | |
| *** yamamoto has quit IRC | 11:14 | |
| *** coolsvap is now known as coolsvap|afk | 11:15 | |
| *** jprovazn has quit IRC | 11:16 | |
| *** ebagdasa_ has quit IRC | 11:16 | |
| *** jprovazn has joined #openstack-neutron | 11:19 | |
| *** popw has quit IRC | 11:24 | |
| *** yog__ has joined #openstack-neutron | 11:28 | |
| *** aepifanov has joined #openstack-neutron | 11:31 | |
| *** trinaths has quit IRC | 11:33 | |
| openstackgerrit | Pavel Bondar proposed openstack/neutron: IPAM reference driver https://review.openstack.org/150485 | 11:40 |
| jlibosva | pc_m: hi | 11:41 |
| *** pal has quit IRC | 11:41 | |
| jlibosva | pc_m: have you started on https://bugs.launchpad.net/neutron/+bug/1441501 yet? I tried to look on that problem and I have a fix ready | 11:41 |
| openstack | Launchpad bug 1441501 in neutron "gate-neutron-fwaas-python27 fails with ImportError: cannot import name test_db_plugin" [Critical,New] - Assigned to Paul Michali (pcm) | 11:41 |
| *** anilvenkata has quit IRC | 11:42 | |
| *** aix has joined #openstack-neutron | 11:42 | |
| *** johnchalekson has joined #openstack-neutron | 11:42 | |
| *** pbrooko has joined #openstack-neutron | 11:42 | |
| openstackgerrit | Jakub Libosvar proposed openstack/neutron-fwaas: Fix import paths after UT cleanup in Neutron https://review.openstack.org/171584 | 11:43 |
| *** ociuhandu has quit IRC | 11:43 | |
| *** johnchalekson has joined #openstack-neutron | 11:47 | |
| *** annegentle has joined #openstack-neutron | 11:54 | |
| openstackgerrit | enikanorov proposed openstack/neutron: Avoid synchronizing session when deleting networkdhcpagentbinding https://review.openstack.org/171588 | 11:55 |
| *** ociuhandu has joined #openstack-neutron | 11:57 | |
| *** yfauser has joined #openstack-neutron | 11:59 | |
| *** yfauser has left #openstack-neutron | 11:59 | |
| *** annegentle has quit IRC | 11:59 | |
| openstackgerrit | enikanorov proposed openstack/neutron: OOP cleanup: start protected method names with underscore https://review.openstack.org/168056 | 11:59 |
| *** yamamoto has joined #openstack-neutron | 12:02 | |
| *** claudiub has joined #openstack-neutron | 12:02 | |
| *** Tiancheng has quit IRC | 12:02 | |
| openstackgerrit | Pavel Bondar proposed openstack/neutron: Implement IPAM Driver loader https://review.openstack.org/147479 | 12:03 |
| openstackgerrit | Ann Kamyshnikova proposed openstack/neutron: Test https://review.openstack.org/171592 | 12:06 |
| *** rotbeard has joined #openstack-neutron | 12:12 | |
| *** anilvenkata has joined #openstack-neutron | 12:13 | |
| *** pc_m_ has joined #openstack-neutron | 12:16 | |
| *** baoli has joined #openstack-neutron | 12:17 | |
| *** pbrooko has quit IRC | 12:18 | |
| *** neeti has quit IRC | 12:20 | |
| *** smccully has joined #openstack-neutron | 12:20 | |
| *** baoli has quit IRC | 12:22 | |
| openstackgerrit | Kevin Benton proposed openstack/neutron: Add simple ARP spoofing protection https://review.openstack.org/171003 | 12:22 |
| *** ociuhandu has quit IRC | 12:22 | |
| *** boris-42 has joined #openstack-neutron | 12:23 | |
| *** baoli has joined #openstack-neutron | 12:25 | |
| *** pbrooko has joined #openstack-neutron | 12:29 | |
| *** baoli_ has joined #openstack-neutron | 12:30 | |
| *** lazy_prince has quit IRC | 12:32 | |
| *** baoli has quit IRC | 12:33 | |
| *** raildo|away is now known as raildo | 12:34 | |
| *** chandankumar has quit IRC | 12:34 | |
| *** yamamoto has quit IRC | 12:37 | |
| *** neeti has joined #openstack-neutron | 12:39 | |
| *** prad has quit IRC | 12:41 | |
| *** amrith has joined #openstack-neutron | 12:43 | |
| openstackgerrit | Paul Michali proposed openstack/neutron-fwaas: FWaaS: Reorganize unit test tree https://review.openstack.org/171602 | 12:43 |
| *** shwetaap has joined #openstack-neutron | 12:44 | |
| *** yatin has quit IRC | 12:46 | |
| *** dave-mccowan has joined #openstack-neutron | 12:47 | |
| openstackgerrit | Romil Gupta proposed openstack/neutron: Added config variable for External Network type in ML2 https://review.openstack.org/149546 | 12:50 |
| *** anilvenkata has quit IRC | 12:53 | |
| *** blues-man has joined #openstack-neutron | 12:54 | |
| *** jpena has joined #openstack-neutron | 12:54 | |
| *** blues-man has quit IRC | 12:54 | |
| *** blues-man has joined #openstack-neutron | 12:54 | |
| *** numan has quit IRC | 12:55 | |
| *** annegentle has joined #openstack-neutron | 12:55 | |
| *** mwagner_lap has joined #openstack-neutron | 12:55 | |
| *** johnchalekson has quit IRC | 12:55 | |
| *** pc_m_ has quit IRC | 12:59 | |
| *** pbrooko has quit IRC | 13:01 | |
| *** johnchalekson has joined #openstack-neutron | 13:02 | |
| *** suresh12 has joined #openstack-neutron | 13:03 | |
| *** jckasper has joined #openstack-neutron | 13:06 | |
| *** pbrooko has joined #openstack-neutron | 13:06 | |
| mestery | pc_m: Thanks for jumping in to fix FWaaS with this one! https://review.openstack.org/#/c/171602/ | 13:07 |
| *** johnbelamaric has joined #openstack-neutron | 13:08 | |
| *** johnchalekson has joined #openstack-neutron | 13:08 | |
| *** numan has joined #openstack-neutron | 13:08 | |
| *** johnchalekson has joined #openstack-neutron | 13:09 | |
| *** asrangne has joined #openstack-neutron | 13:09 | |
| *** johnchalekson has joined #openstack-neutron | 13:10 | |
| *** julim has joined #openstack-neutron | 13:11 | |
| *** anilvenkata has joined #openstack-neutron | 13:11 | |
| *** johnchalekson has joined #openstack-neutron | 13:11 | |
| *** tomoe_ has joined #openstack-neutron | 13:11 | |
| *** aswadr has quit IRC | 13:12 | |
| *** rpothier has joined #openstack-neutron | 13:12 | |
| *** johnchalekson has joined #openstack-neutron | 13:12 | |
| *** johnchalekson has joined #openstack-neutron | 13:13 | |
| *** amrith is now known as _amrith_ | 13:14 | |
| *** tiswanso has joined #openstack-neutron | 13:15 | |
| *** tomoe_ has quit IRC | 13:16 | |
| *** lukasa has joined #openstack-neutron | 13:16 | |
| *** _amrith_ is now known as amrith | 13:16 | |
| *** moshele has quit IRC | 13:16 | |
| *** roeyc has joined #openstack-neutron | 13:16 | |
| mestery | HenryG: I see dane_leblanc opened https://bugs.launchpad.net/neutron/+bug/1441382, I targeted it at RC!, but I must admit, I need help in the strategy for solving this one and https://bugs.launchpad.net/neutron/+bug/1440183 by tomorrow! :) | 13:17 |
| openstack | Launchpad bug 1441382 in neutron "IPv6 SLAAC subnet Tempest tests fail due to IntegrityError " [High,New] - Assigned to Dane LeBlanc (leblancd) | 13:17 |
| openstack | Launchpad bug 1440183 in neutron "DBDeadlock on subnet allocation" [Critical,In progress] - Assigned to Dane LeBlanc (leblancd) | 13:17 |
| openstackgerrit | Pavel Bondar proposed openstack/neutron: Refactor db_base_plugin_v2.py to use new IPAM https://review.openstack.org/153236 | 13:17 |
| *** yamahata has joined #openstack-neutron | 13:17 | |
| openstackgerrit | Kyle Mestery proposed openstack/neutron-fwaas: FWaaS: Reorganize unit test tree https://review.openstack.org/171602 | 13:20 |
| *** ihrachyshka has left #openstack-neutron | 13:21 | |
| *** ihrachyshka has joined #openstack-neutron | 13:21 | |
| *** Longgeek has quit IRC | 13:23 | |
| *** ivar-lazzaro has joined #openstack-neutron | 13:24 | |
| *** ociuhandu has joined #openstack-neutron | 13:24 | |
| *** jpena has quit IRC | 13:24 | |
| *** jpena has joined #openstack-neutron | 13:25 | |
| *** ivar-laz_ has joined #openstack-neutron | 13:28 | |
| *** ivar-lazzaro has quit IRC | 13:29 | |
| *** ivar-laz_ has quit IRC | 13:30 | |
| *** ivar-lazzaro has joined #openstack-neutron | 13:30 | |
| *** pc_m_ has joined #openstack-neutron | 13:31 | |
| openstackgerrit | Paul Michali proposed openstack/neutron-fwaas: FWaaS: Reorganize unit test tree https://review.openstack.org/171602 | 13:32 |
| *** pc_m_ has quit IRC | 13:32 | |
| *** thomasem has joined #openstack-neutron | 13:34 | |
| mestery | pc_m: Thanks for the 3rd patch on that FWaaS UT fixup | 13:34 |
| *** anilvenkata has quit IRC | 13:35 | |
| *** zhenq has joined #openstack-neutron | 13:35 | |
| *** thomasem has quit IRC | 13:36 | |
| *** amuller has joined #openstack-neutron | 13:37 | |
| pc_m | mestery: Sure, didn't see anyone on it, so I jumped on it. | 13:37 |
| pc_m | mestery: Will follow up with the bash change too for FW. | 13:38 |
| *** rushil has joined #openstack-neutron | 13:39 | |
| *** Marga_ has joined #openstack-neutron | 13:39 | |
| *** dane_leblanc has joined #openstack-neutron | 13:40 | |
| *** Marga_ has quit IRC | 13:40 | |
| *** haleyb_ has joined #openstack-neutron | 13:40 | |
| *** Marga_ has joined #openstack-neutron | 13:40 | |
| *** thomasem has joined #openstack-neutron | 13:41 | |
| *** johnchalekson has quit IRC | 13:43 | |
| *** johnchalekson has joined #openstack-neutron | 13:44 | |
| *** johnchalekson has quit IRC | 13:44 | |
| *** kbringard has joined #openstack-neutron | 13:44 | |
| *** numan has quit IRC | 13:44 | |
| *** johnchalekson has joined #openstack-neutron | 13:45 | |
| openstackgerrit | Paul Michali proposed openstack/neutron-fwaas: FWaaS: Remove check for bash usage https://review.openstack.org/171637 | 13:45 |
| *** pal has joined #openstack-neutron | 13:45 | |
| *** jaypipes has joined #openstack-neutron | 13:48 | |
| *** apuimedo has quit IRC | 13:50 | |
| pc_m | mestery: Here's the companion commit to remove bash check ^^ | 13:50 |
| mestery | pc_m: I -1'd that already, I'n not holding the release for a bash check fix :) | 13:50 |
| pc_m | mestery: I created bug 1441632 | 13:50 |
| openstack | bug 1441632 in neutron "FWaaS: Remove checks for bash scripts" [Low,In progress] https://launchpad.net/bugs/1441632 - Assigned to Paul Michali (pcm) | 13:50 |
| mestery | pc_m: We have too many bugs, and we seem to keep adding more | 13:51 |
| mestery | We're never going to get the release at this rate :) | 13:51 |
| pc_m | mestery: Sure. np. | 13:51 |
| mestery | And I don't think bash removal is a release blocker | 13:51 |
| mestery | So, Liberty it is! | 13:51 |
| mestery | pc_m: But thanks for all your hard work! You're an inspiration :) | 13:51 |
| *** pal has quit IRC | 13:52 | |
| *** anilvenkata has joined #openstack-neutron | 13:52 | |
| pc_m | mestery: np. I had scripting to help do this for VPN, so it was easy to reuse for FW. | 13:52 |
| *** dims has quit IRC | 13:52 | |
| mestery | pc_m: nice! | 13:52 |
| *** dims has joined #openstack-neutron | 13:53 | |
| *** jistr has quit IRC | 13:54 | |
| *** jistr has joined #openstack-neutron | 13:54 | |
| *** yamahata has quit IRC | 13:55 | |
| *** yamahata has joined #openstack-neutron | 13:55 | |
| *** aix has quit IRC | 13:56 | |
| pc_m | mestery: So what is the process for this release? Will there be another RC, or will this one be it? When does Liberty (master) open up? | 13:56 |
| *** dboik has joined #openstack-neutron | 13:58 | |
| *** chlong has quit IRC | 13:58 | |
| *** pdb has quit IRC | 13:59 | |
| openstackgerrit | enikanorov proposed openstack/neutron: Add logging to dangling port to ml2 delete_subnet https://review.openstack.org/171662 | 13:59 |
| openstackgerrit | enikanorov proposed openstack/neutron: Add logging of dangling port to ml2 delete_subnet https://review.openstack.org/171662 | 14:00 |
| *** chandankumar has joined #openstack-neutron | 14:00 | |
| *** nplanel_ has joined #openstack-neutron | 14:01 | |
| *** jobewan has joined #openstack-neutron | 14:03 | |
| *** anish_ has quit IRC | 14:03 | |
| *** numan has joined #openstack-neutron | 14:04 | |
| *** killer_prince has joined #openstack-neutron | 14:05 | |
| *** oomichi has quit IRC | 14:05 | |
| *** killer_prince is now known as lazy_prince | 14:05 | |
| *** ivar-lazzaro has quit IRC | 14:08 | |
| *** haleyb_ has quit IRC | 14:09 | |
| *** jistr has quit IRC | 14:10 | |
| *** ociuhandu has quit IRC | 14:11 | |
| *** ociuhandu_ has joined #openstack-neutron | 14:11 | |
| *** moshele has joined #openstack-neutron | 14:12 | |
| *** salv-orlando has quit IRC | 14:12 | |
| *** jistr has joined #openstack-neutron | 14:12 | |
| *** zz_jgrimm is now known as jgrimm | 14:13 | |
| *** suresh12 has quit IRC | 14:15 | |
| openstackgerrit | Ann Kamyshnikova proposed openstack/neutron: Add additional time for starting l3 agents during scheduling https://review.openstack.org/171592 | 14:17 |
| openstackgerrit | Paul Michali proposed openstack/neutron-fwaas: FWaaS: Reorganize unit test tree https://review.openstack.org/171602 | 14:18 |
| mestery | pc_m: I hope there isn't an RC2, but you never know. | 14:18 |
| mestery | pc_m: The process is the same as it always is: Once we cut the RC branch I'll merge the patches opening up Liberty, and the master branch will then be open | 14:18 |
| mestery | pc_m: Any new RC issues will be cherry-picked over to the RC branch for a potential RC2 | 14:19 |
| mestery | pc_m: And then, I can finally rest :) | 14:19 |
| *** leenheer has joined #openstack-neutron | 14:19 | |
| *** asrangne__ has joined #openstack-neutron | 14:20 | |
| pc_m | mestery: Thanks. I couldn't recall the steps... so if this RC1 looks ok, there will be a merge period and then we can update patches targeted toward liberty and continue on reviews. Gotcha. | 14:20 |
| ihrachyshka | dougwig, hey. a question re neutron_*aas.conf files | 14:20 |
| mestery | pc_m: Once we cut the RC, we'll open master and it becomes Liberty at that time :) | 14:21 |
| pc_m | mestery: Should I change reviews I have to WIP? | 14:21 |
| ihrachyshka | dougwig, wasn't it supposed that we load them magically into neutron-server without passing --config-file args? | 14:21 |
| ihrachyshka | dougwig, or is magic applied for specific sections and cases only? | 14:21 |
| mestery | pc_m: Nope, I assume cores are paying attention and not merging stuff this late in the release unless it's on the RC1 list | 14:21 |
| openstackgerrit | Arvind Somya proposed openstack/neutron: Model and migration for cisco providernet extension driver https://review.openstack.org/171326 | 14:21 |
| mestery | ihrachyshka: dougwig is on the west coast this week and may not be awake yet, just FYI :) | 14:22 |
| *** pbrooko has quit IRC | 14:22 | |
| *** asrangne has quit IRC | 14:22 | |
| ihrachyshka | mestery, ah ok, thanks | 14:22 |
| mestery | ihrachyshka: yw | 14:22 |
| openstackgerrit | enikanorov proposed openstack/neutron: Add logging to dangling port to ml2 delete_subnet https://review.openstack.org/171662 | 14:23 |
| *** vivek-ebay has joined #openstack-neutron | 14:24 | |
| *** aix has joined #openstack-neutron | 14:24 | |
| *** Sukhdev has joined #openstack-neutron | 14:24 | |
| *** yamamoto has joined #openstack-neutron | 14:26 | |
| openstackgerrit | Ann Kamyshnikova proposed openstack/neutron: Add additional time for starting l3 agents during scheduling https://review.openstack.org/171592 | 14:26 |
| *** kiran-r has quit IRC | 14:27 | |
| *** ivar-lazzaro has joined #openstack-neutron | 14:27 | |
| *** zzzeek has joined #openstack-neutron | 14:28 | |
| *** prad has joined #openstack-neutron | 14:28 | |
| *** anilvenkata has quit IRC | 14:29 | |
| *** yamamoto has quit IRC | 14:31 | |
| *** chandankumar has quit IRC | 14:33 | |
| *** vivek-eb_ has joined #openstack-neutron | 14:34 | |
| *** pradipta has joined #openstack-neutron | 14:34 | |
| openstackgerrit | Andrew Boik proposed openstack/neutron: Consider all address scopes in init_l3 https://review.openstack.org/168502 | 14:36 |
| *** vivek-ebay has quit IRC | 14:36 | |
| openstackgerrit | Merged openstack/python-neutronclient: Make secgroup rules more readable in security-group-show https://review.openstack.org/160222 | 14:37 |
| openstackgerrit | Merged openstack/python-neutronclient: Update hacking to 0.10 https://review.openstack.org/166702 | 14:38 |
| *** neiljerram has joined #openstack-neutron | 14:39 | |
| openstackgerrit | Merged openstack/python-neutronclient: Remove author tag https://review.openstack.org/166714 | 14:39 |
| neiljerram | enikanorov_: ping | 14:39 |
| openstackgerrit | Merged openstack/python-neutronclient: Fix E265 block comment should start with '# ' https://review.openstack.org/166715 | 14:41 |
| *** aepifanov has quit IRC | 14:42 | |
| *** aepifanov has joined #openstack-neutron | 14:42 | |
| *** aepifanov has quit IRC | 14:43 | |
| *** aepifanov1 has joined #openstack-neutron | 14:43 | |
| openstackgerrit | Nikolay Fedotov proposed openstack/neutron-vpnaas: Scenario test for vpnaas: ipsec-site-connection https://review.openstack.org/159746 | 14:43 |
| *** ihrachyshka has quit IRC | 14:44 | |
| *** tomoe_ has joined #openstack-neutron | 14:44 | |
| *** jistr has quit IRC | 14:45 | |
| mestery | dane_leblanc: What is going on with this bug https://bugs.launchpad.net/neutron/+bug/1441382 | 14:46 |
| openstack | Launchpad bug 1441382 in neutron "IPv6 SLAAC subnet Tempest tests fail due to IntegrityError " [High,New] - Assigned to Dane LeBlanc (leblancd) | 14:46 |
| mestery | dane_leblanc: Are you planning to fix this? Is this an RC blocker? | 14:46 |
| *** emagana has joined #openstack-neutron | 14:46 | |
| *** moshele has quit IRC | 14:46 | |
| dane_leblanc | mestery: Yes, I'm planning to fix this. I have something that appears to work (with Salvatore's help), but need some more soak time. | 14:47 |
| *** neiljerram has quit IRC | 14:47 | |
| *** neiljerram has joined #openstack-neutron | 14:47 | |
| *** carl_baldwin has joined #openstack-neutron | 14:48 | |
| dane_leblanc | mestery: I don't think the integrity error is as frequent as the deadlock timeout was, but I still think it's frequent enough to be a serious issue. | 14:48 |
| mestery | dane_leblanc: In your judgement, should we block the release on that bug? Doesn't sound like it to me | 14:49 |
| *** neiljerram has quit IRC | 14:49 | |
| *** neiljerram has joined #openstack-neutron | 14:49 | |
| *** emagana_ has joined #openstack-neutron | 14:49 | |
| enikanorov_ | neiljerram: pong | 14:49 |
| *** emagana has quit IRC | 14:50 | |
| neiljerram | Ah, hi there. | 14:50 |
| enikanorov_ | hi | 14:50 |
| *** Murali has joined #openstack-neutron | 14:50 | |
| dane_leblanc | mestery: It depends on exactly how frequent the failures happen in the gate, but so far the deadlock timeout (the outer layer of the onion) happens instead, so we don't have a perfect view of what the occurence is. | 14:50 |
| *** claudiub has quit IRC | 14:50 | |
| mestery | dane_leblanc: Ack, I'll leave it there for now and hopefully we can close on the DB layer one first. | 14:51 |
| mestery | dane_leblanc: Which review are you targeting for the DB layer fix? | 14:51 |
| mestery | carl_baldwin: I +A'd this review, but it's not going in the gate, is it dependent on something I'm not aware of? https://review.openstack.org/#/c/154289/ | 14:51 |
| *** neiljerram has quit IRC | 14:51 | |
| *** jistr has joined #openstack-neutron | 14:51 | |
| mestery | carl_baldwin: Your comment near the end made me wonder | 14:51 |
| *** neiljerram has joined #openstack-neutron | 14:52 | |
| mestery | carl_baldwin: Nevermind, figured it out. | 14:52 |
| carl_baldwin | mestery, :) | 14:52 |
| neiljerram | enikanorov_: sorry, just had a disconnect, trying again... | 14:53 |
| carl_baldwin | mestery, I had an explanation partly written. Others in channel might be curious. It is the Depends-On line in the commit message holding it up. | 14:53 |
| mestery | carl_baldwin: Sorry, it's been a long few weeks, I think I'm losing what's left of my mind | 14:53 |
| dane_leblanc | mestery: DB layer? You mean the DB deadlock timeout? That will be addressed with https://review.openstack.org/#/c/170690/, although no one is clear on *why* this fixes the problem | 14:53 |
| mestery | carl_baldwin: Yes, working with mtreinish to try and get QA folks to merge that one now. | 14:53 |
| neiljerram | enikanorov_: I was just wondering whether you might have further info on https://bugs.launchpad.net/neutron/+bug/1381536 | 14:53 |
| openstack | Launchpad bug 1381536 in neutron "ResourceClosedError occurs when neutron API run in parallel" [High,Confirmed] - Assigned to Neil Jerram (neil-jerram) | 14:53 |
| mestery | dane_leblanc: Yes, db deadlock | 14:53 |
| carl_baldwin | mestery, good thing it worked. Otherwise, we'd have a broken gate. Nice new little feature. | 14:54 |
| dane_leblanc | mestery: But I can reproduce the problem locally, and it sure does seem to fix it. | 14:54 |
| neiljerram | enikanorov_: It's not a particularly important bug, right now, but it would be nice to keep nudging it along. | 14:54 |
| mestery | carl_baldwin: lol | 14:54 |
| *** phil_h has joined #openstack-neutron | 14:54 | |
| mestery | dane_leblanc: OK, lets see if we can merge it, I see rossella_s had some comments on that, you shoudl look at those now. | 14:54 |
| enikanorov_ | neiljerram: i agree, but my gut feeling tells me that we might not see any evidence of this issue again | 14:55 |
| mestery | carl_baldwin: We got the tempest one in the queue now, so fingers crossed, the neutron side can land soon too | 14:55 |
| dane_leblanc | mestery: Right-o. Will make that change. | 14:56 |
| neiljerram | enikanorov_: You suspect it's been fixed by some other change, since it was reported? | 14:56 |
| enikanorov_ | neiljerram: yes, at least for ml2 plugin | 14:56 |
| mestery | dane_leblanc: Excellent sir! | 14:56 |
| amotoki | mestery: what is your opinion on my comment in https://review.openstack.org/#/c/169569/18? By default all mechanism drivers except OVS are assumed to support VLAN transparency. | 14:57 |
| amotoki | mestery: pls check after other stuffs. you have many stuffs! | 14:57 |
| mestery | amotoki: lol, looking now | 14:57 |
| neiljerram | enikanorov_: OK, cool. Should we somehow mark or expire the bug then? | 14:57 |
| *** jobewan has quit IRC | 14:57 | |
| *** rangansw has joined #openstack-neutron | 14:57 | |
| neiljerram | enikanorov_: Or maybe just leave it sitting for a while, and then it will auto-expire? | 14:58 |
| enikanorov_ | neiljerram: usually it's done by marking it incomplete | 14:58 |
| mestery | amotoki: I agree with you, I changed my vote and once Pritesh fixes that, we can merge it I think,. | 14:59 |
| mestery | amotoki: If you want to roll a new patch with that change, I am good with that. | 14:59 |
| mestery | amotoki: Up to you | 14:59 |
| neiljerram | enikanorov_: OK, fine, I will do that. Many thanks! | 14:59 |
| amotoki | mestery: I am fine with your idea. I will file a bug and propose the fix. | 14:59 |
| *** pal has joined #openstack-neutron | 15:00 | |
| mestery | amotoki: What is the bug for? | 15:00 |
| mestery | amotoki: You can just re-roll pritech's patch and swap the default per your comment. | 15:00 |
| openstackgerrit | Baodong (Robert) Li proposed openstack/neutron: Implement Neutron IPv6 Prefix Delegation https://review.openstack.org/158697 | 15:00 |
| openstackgerrit | Nikolay Fedotov proposed openstack/neutron-vpnaas: Scenario test for vpnaas: ipsec-site-connection https://review.openstack.org/159746 | 15:00 |
| amotoki | mestery: ooo... I missed you. you are waiting for the fix. | 15:01 |
| *** busterswt has joined #openstack-neutron | 15:02 | |
| *** rangansw has quit IRC | 15:02 | |
| *** jaypipes has quit IRC | 15:03 | |
| amotoki | mestery: I changed my score too. | 15:03 |
| openstackgerrit | enikanorov proposed openstack/neutron: Add logging to dangling port to ml2 delete_subnet https://review.openstack.org/171662 | 15:03 |
| *** emagana_ has quit IRC | 15:05 | |
| *** neiljerram has quit IRC | 15:05 | |
| *** jaypipes has joined #openstack-neutron | 15:05 | |
| *** emagana has joined #openstack-neutron | 15:06 | |
| *** Marga_ has quit IRC | 15:06 | |
| *** jaypipes has quit IRC | 15:06 | |
| *** chandankumar has joined #openstack-neutron | 15:06 | |
| mestery | amotoki: Cool, thanks! | 15:06 |
| mestery | amotoki: I will get back to you on the client release email you sent later today | 15:07 |
| *** xgerman has joined #openstack-neutron | 15:07 | |
| amotoki | mestery: thanks. Perhaps we need to discuss the version cap in the dev list more broadly to make things smooth. | 15:09 |
| openstackgerrit | Merged openstack/neutron: linuxbridge UT: Fix a regression of the recent ip_lib change https://review.openstack.org/165291 | 15:09 |
| mestery | amotoki: Sure, I'll digest it as well now and respond. | 15:09 |
| *** xgerman has quit IRC | 15:09 | |
| *** xgerman has joined #openstack-neutron | 15:10 | |
| *** pal has quit IRC | 15:11 | |
| *** salv-orlando has joined #openstack-neutron | 15:12 | |
| *** abregman has quit IRC | 15:13 | |
| openstackgerrit | Dane LeBlanc proposed openstack/neutron: Re-use context session in ML2 DB get_port_binding_host https://review.openstack.org/170690 | 15:14 |
| *** ronis has quit IRC | 15:15 | |
| *** jistr is now known as jistr|mtg | 15:16 | |
| *** moshele has joined #openstack-neutron | 15:16 | |
| *** pal has joined #openstack-neutron | 15:17 | |
| openstackgerrit | Miguel Angel Ajo proposed openstack/neutron: (DO NOT MERGE/WIP/POC) OVSFirewallDriver https://review.openstack.org/167671 | 15:18 |
| *** _cjones_ has joined #openstack-neutron | 15:20 | |
| *** pritesh has joined #openstack-neutron | 15:23 | |
| openstackgerrit | Pavel Bondar proposed openstack/neutron: IPAM reference driver https://review.openstack.org/150485 | 15:23 |
| *** vivek-eb_ has quit IRC | 15:23 | |
| *** lazy_prince has quit IRC | 15:24 | |
| mestery | pritesh: See comments on your VLAN patch, if you re-roll it quickly, amotoki and I can merge it for you. thanks! | 15:24 |
| *** _cjones_ has quit IRC | 15:24 | |
| *** banix has joined #openstack-neutron | 15:24 | |
| *** armax has joined #openstack-neutron | 15:25 | |
| *** marrusl has quit IRC | 15:26 | |
| *** suresh12 has joined #openstack-neutron | 15:26 | |
| *** yog__ has quit IRC | 15:26 | |
| *** numan has quit IRC | 15:27 | |
| *** sudipto has joined #openstack-neutron | 15:28 | |
| *** killer_prince has joined #openstack-neutron | 15:29 | |
| pritesh | mestery: already on it :) | 15:29 |
| mestery | pritesh: Thanks man :) | 15:29 |
| *** killer_prince is now known as lazy_prince | 15:29 | |
| *** dboik_ has joined #openstack-neutron | 15:29 | |
| pritesh | mestery: no worries, thanks to you for helping :) | 15:29 |
| mestery | pritesh: :) | 15:30 |
| *** suresh12 has quit IRC | 15:30 | |
| *** jistr|mtg is now known as jistr | 15:30 | |
| openstackgerrit | pritesh proposed openstack/neutron: Moving VLAN Transparency support from core to extension https://review.openstack.org/169569 | 15:31 |
| pritesh | mestery: actually rolling out another patch, found another small change in comment | 15:32 |
| mestery | pritesh: Is it included in the one ^^^^ ??? | 15:32 |
| *** sudipto_ has joined #openstack-neutron | 15:32 | |
| *** dboik has quit IRC | 15:33 | |
| pritesh | mestery: no but is included in this next one below | 15:33 |
| *** shwetaap has quit IRC | 15:33 | |
| mestery | pritesh: Ack | 15:33 |
| *** bobmel_ has joined #openstack-neutron | 15:34 | |
| *** nlahouti has joined #openstack-neutron | 15:34 | |
| *** thedodd has joined #openstack-neutron | 15:34 | |
| mestery | amotoki: pritesh has made the change, please re-review! | 15:34 |
| amotoki | mestery: Iooking already :) | 15:35 |
| openstackgerrit | pritesh proposed openstack/neutron: Moving VLAN Transparency support from core to extension https://review.openstack.org/169569 | 15:35 |
| mestery | carl_baldwin: I have cherry-picked this client patch (https://review.openstack.org/#/c/159618/), but it's not clear to me how to create a subnet from the subnetpool I've created. Am I missing something or do we need another CLI patch which allows passing a subnetpool ID into subnet-create? | 15:35 |
| *** matrohon has quit IRC | 15:35 | |
| pritesh | sorry guys, the help text was old, had to change it to reflect the current exception it raises, rest is same | 15:35 |
| *** lazy_prince has quit IRC | 15:36 | |
| amotoki | pritesh: thanks for your quick update. one request: please remove vlan_transparency from the sample neturon.conf. | 15:36 |
| *** claudiub has joined #openstack-neutron | 15:36 | |
| *** rpothier has quit IRC | 15:36 | |
| *** bobmel has quit IRC | 15:36 | |
| amotoki | pritesh: otherwise looks good perfect. | 15:36 |
| pritesh | amotoki: but we need config option to turn it on or off, so not sure how people will know if it exists ? | 15:37 |
| * carl_baldwin opening patch | 15:37 | |
| mestery | carl_baldwin: I'm sorry to be bothering you so much this morning :( | 15:37 |
| *** rwsu has joined #openstack-neutron | 15:37 | |
| *** tsg has joined #openstack-neutron | 15:37 | |
| amotoki | peterstac: I think you remove the option in ML2 plugin and it was used only in ML2 plugin. Am I missing something? | 15:37 |
| carl_baldwin | mystery: Do you have https://review.openstack.org/#/c/165910/ ? | 15:38 |
| openstackgerrit | Edgar Magana proposed openstack/neutron: Add missing config parameters in neutron.conf https://review.openstack.org/171059 | 15:38 |
| * mestery looks | 15:38 | |
| pritesh | amotoki: https://review.openstack.org/#/c/169569/20/neutron/extensions/vlantransparent.py#42 | 15:39 |
| mestery | carl_baldwin: Since I did a "git review -d", I bet not :( | 15:39 |
| mestery | carl_baldwin: Let me cherry-pick it on top instead and see if that helps, thanks! | 15:39 |
| *** rpothier has joined #openstack-neutron | 15:39 | |
| amotoki | peterstac: ah. got it! | 15:39 |
| openstackgerrit | Edgar Magana proposed openstack/neutron: Add missing config parameters in neutron.conf https://review.openstack.org/171059 | 15:40 |
| pritesh | amotoki: ok thanks :) hopefully now everything looks good? | 15:40 |
| carl_baldwin | mestery, it adds the --subnetpool and --prefixlen options which you would provide in place of a --cidr. | 15:40 |
| mestery | carl_baldwin: Yes, I'm doing it now | 15:40 |
| amotoki | pritesh: now checking as the final round :) | 15:40 |
| pritesh | amotoki: sounds good, thanks for your and mestery's help. | 15:41 |
| *** afazekas has quit IRC | 15:43 | |
| openstackgerrit | Dane LeBlanc proposed openstack/neutron: Catch and ignore integrity errors for subnet create https://review.openstack.org/170968 | 15:43 |
| *** ozialien has joined #openstack-neutron | 15:44 | |
| openstackgerrit | Sridhar Gaddam proposed openstack/neutron: Neutron to Drop Router Advts from VM ports https://review.openstack.org/140046 | 15:45 |
| *** srampal has joined #openstack-neutron | 15:45 | |
| *** ozialien has quit IRC | 15:47 | |
| *** SridharG1 has left #openstack-neutron | 15:49 | |
| *** alex_xu has quit IRC | 15:50 | |
| *** alex_xu has joined #openstack-neutron | 15:50 | |
| mestery | carl_baldwin: Tahnks for the pointer, the problem was I did a "git review -d" for the subnetpool CRUD stuff in the client, and since it wasn't based on the review you mentioned, I lost that part. | 15:50 |
| mestery | carl_baldwin: Once I directly cherry-picked it, things worked out much better. | 15:50 |
| *** shwetaap has joined #openstack-neutron | 15:52 | |
| *** dannywilson has joined #openstack-neutron | 15:52 | |
| *** doude has quit IRC | 15:53 | |
| *** gsagie_ has joined #openstack-neutron | 15:53 | |
| *** chandankumar has quit IRC | 15:55 | |
| *** killer_prince has joined #openstack-neutron | 15:55 | |
| emagana | anteaya: hello! | 15:55 |
| *** killer_prince is now known as lazy_prince | 15:55 | |
| emagana | anteaya: In case you missed but the nova-net to neutron initial doc was merged as .rst instead of xml: https://review.openstack.org/#/c/167076/2 | 15:56 |
| *** jistr has quit IRC | 15:56 | |
| *** neeti has quit IRC | 15:56 | |
| *** kfox1111 has joined #openstack-neutron | 15:57 | |
| openstackgerrit | Merged openstack/neutron: Add ipset element and hashsize tunables https://review.openstack.org/170328 | 15:57 |
| *** Marga_ has joined #openstack-neutron | 15:57 | |
| mestery | carl_baldwin: Subnetpool allocation is so awesome, thanks to you and ryan for getting this in! It's a fantastic feature! :) | 15:57 |
| kfox1111 | Is any of the l3 router developers online? | 15:57 |
| *** _cjones_ has joined #openstack-neutron | 15:57 | |
| mestery | haleyb: ipset finally merged! Woot! ^^^^ | 15:57 |
| mestery | kfox1111: carl_baldwin is here for you, and also amuller :) | 15:57 |
| *** anshul has quit IRC | 15:57 | |
| *** yamahata has quit IRC | 15:58 | |
| *** yamahata has joined #openstack-neutron | 15:58 | |
| *** cbouch has joined #openstack-neutron | 15:58 | |
| *** kiran has joined #openstack-neutron | 15:58 | |
| kfox1111 | Vikram and I are talking about adding ntp time support to Neutron. We're debating if we should extend the router interface to support it, since a lot of hardware routers support ntp natively. Or should we make it a totally seperate service, more like dhcp? | 15:59 |
| *** jlibosva has quit IRC | 15:59 | |
| *** marun has joined #openstack-neutron | 15:59 | |
| *** ozialien has joined #openstack-neutron | 15:59 | |
| *** Marga_ has quit IRC | 16:00 | |
| *** Marga_ has joined #openstack-neutron | 16:00 | |
| *** igordcard has quit IRC | 16:00 | |
| kfox1111 | My preference is to add it to the router since dvr would allow the tenant ntp servers right on the compute nodes, scaling very well. But it is adding more complexity to the router. | 16:01 |
| *** igordcard has joined #openstack-neutron | 16:01 | |
| *** ozialien has quit IRC | 16:01 | |
| kfox1111 | Is that unpalatable? | 16:01 |
| sc68cal | kfox1111: what does this get us compared to having most guest images point to pool.ntp.org | 16:01 |
| *** aepifanov1 has quit IRC | 16:02 | |
| kfox1111 | we don't beat pool.ntp.org down? :) | 16:02 |
| *** absubram has joined #openstack-neutron | 16:02 | |
| sc68cal | kfox1111: pool.ntp.org is an anycast address | 16:02 |
| *** aepifanov has joined #openstack-neutron | 16:02 | |
| openstackgerrit | Kevin Benton proposed openstack/neutron: Add simple ARP spoofing protection https://review.openstack.org/171003 | 16:02 |
| kfox1111 | as far as I understand it, the default ubuntu image defaults into ubuntu's pool, and centos into cento's, and rhel into redhats... | 16:03 |
| kfox1111 | I dont think any uses pool.ntp.org directly. | 16:03 |
| *** dboik_ has quit IRC | 16:03 | |
| sc68cal | so then we'd need to update images to use an ntp service that neutron would offer, yes? | 16:04 |
| kfox1111 | As a cloud provider, I'd like to be a good citizen and provide a local endpoint. | 16:04 |
| *** dboik has joined #openstack-neutron | 16:04 | |
| kfox1111 | yeah. I was thinking, ntp should respond on 169.254.169.254 so that its always constant on any openstack cloud. | 16:04 |
| pritesh | mestery: got blessing from amotoki on #169569, need your blessing on it, hopefully last time before it merges :) | 16:04 |
| *** crose has joined #openstack-neutron | 16:05 | |
| *** ozialien has joined #openstack-neutron | 16:05 | |
| sc68cal | 169.254.169.254 is amazon metadata api compat | 16:05 |
| sc68cal | their service does not provide ntp, unless i am mistaken | 16:05 |
| kfox1111 | yeah. but it only uses port 80. so putting port 123 on it wouldn't be bad? | 16:05 |
| *** ozialien has quit IRC | 16:06 | |
| *** luqas has quit IRC | 16:06 | |
| haleyb | mestery: thanks for the rechecks, i was worried that would just keep failing... | 16:06 |
| kfox1111 | would just take a simple iptables rule in the l3router namespace. | 16:06 |
| *** YorikSar has quit IRC | 16:06 | |
| *** luqas has joined #openstack-neutron | 16:06 | |
| sc68cal | kfox1111: yes, but it would be making a change for something that is not really worthwile in my opinion | 16:06 |
| *** luqas has quit IRC | 16:06 | |
| sc68cal | since you'd need to configure guest images to utilize | 16:06 |
| kfox1111 | what I'm trying to do long term, is eliminate the need to customize the image constantly... | 16:07 |
| kfox1111 | right now, every heat template I write has to tweak ntp.conf, or | 16:07 |
| kfox1111 | I have to rebuild images. | 16:07 |
| kfox1111 | I'd like for 'openstack compatable images' to default to 169.254.169.254 to allow the cloud provider to point it at the right place for that cloud. | 16:08 |
| haleyb | kfox1111: can't you just specify the ntp server in the dhcp reply? option 42 | 16:08 |
| *** chandankumar has joined #openstack-neutron | 16:08 | |
| kfox1111 | Our clouds can point at our corporate ntp pool for example. | 16:08 |
| kfox1111 | hmm... not sure. let me look. | 16:09 |
| *** dave-mccowan has quit IRC | 16:09 | |
| sc68cal | and we already have a way to pass in extra_dhcp_opts | 16:09 |
| haleyb | or maybe you can tweak your dns to respond with a local IP for ntp.pool.org or similar | 16:09 |
| kfox1111 | interesting... | 16:10 |
| kfox1111 | do ubuntu/centos/rhel support that option today? | 16:10 |
| openstackgerrit | Merged openstack/neutron: Return from check_ports_exist_on_l3agent if no subnet found https://review.openstack.org/154289 | 16:11 |
| haleyb | kfox1111: dhclient.conf on 14.04 by default requests ntp-servers | 16:11 |
| *** kiran has quit IRC | 16:12 | |
| kfox1111 | interesting. I'll give it a try with our existing images and see if it just works. Thanks for the pointer. :) | 16:12 |
| *** johnbelamaric has quit IRC | 16:12 | |
| *** dave-mccowan has joined #openstack-neutron | 16:13 | |
| *** julim has quit IRC | 16:14 | |
| kfox1111 | Can you set extra_dhcp_opt per subnet, or is it a neutron.conf thing? | 16:16 |
| anteaya | emagana: I hadn't known that, thanks for the link, and yay .rst | 16:17 |
| kfox1111 | ah. https://blueprints.launchpad.net/neutron/+spec/extra-dhcp-opts-ipv4-ipv6 | 16:17 |
| emagana | anteaya: ;-) | 16:17 |
| *** kiran has joined #openstack-neutron | 16:17 | |
| *** tsg has quit IRC | 16:17 | |
| kfox1111 | so I can't use it yet. but as soon as kilo comes out... | 16:17 |
| anteaya | :) | 16:18 |
| kfox1111 | hmm... no, thats not quite the right blueprint... | 16:18 |
| *** ozialien has joined #openstack-neutron | 16:19 | |
| kfox1111 | wait... is extra_dhcp_opts per port, not on the subnet? | 16:20 |
| carl_baldwin | mestery, thanks. | 16:20 |
| kfox1111 | it would hurt to have to manually do it on all vm's... Is there a subnet eqiv? | 16:20 |
| *** neeti has joined #openstack-neutron | 16:22 | |
| *** ozialien has quit IRC | 16:23 | |
| kfox1111 | So... no. :/ | 16:23 |
| kfox1111 | https://blueprints.launchpad.net/neutron/+spec/dhcp-options-per-subnet | 16:23 |
| *** Sukhdev has quit IRC | 16:24 | |
| *** pask81 has quit IRC | 16:24 | |
| *** Sukhdev has joined #openstack-neutron | 16:25 | |
| *** moshele has quit IRC | 16:25 | |
| *** kiran has quit IRC | 16:25 | |
| *** Longgeek has joined #openstack-neutron | 16:25 | |
| *** ygbo has quit IRC | 16:25 | |
| *** kiran has joined #openstack-neutron | 16:25 | |
| *** moshele has joined #openstack-neutron | 16:25 | |
| *** ijw has joined #openstack-neutron | 16:27 | |
| *** moshele has quit IRC | 16:27 | |
| *** catohornet has quit IRC | 16:27 | |
| carl_baldwin | FWIW, I think I like the DHCP options possibility over adding anything new to the router. But it may still take some work. | 16:27 |
| carl_baldwin | kfox1111: http://support.ntp.org/bin/view/Support/ConfiguringNTP#Section_6.12. | 16:27 |
| *** asrangne__ has quit IRC | 16:28 | |
| carl_baldwin | ^ See the parts about configuring the clients. This would take some work in the images too. But, I think I'd like that over using 169.254... | 16:28 |
| *** claudiub has quit IRC | 16:28 | |
| *** irenab has quit IRC | 16:28 | |
| *** ijw has quit IRC | 16:29 | |
| *** ijw has joined #openstack-neutron | 16:29 | |
| *** lukasa has quit IRC | 16:29 | |
| *** kiran has quit IRC | 16:30 | |
| *** Longgeek has quit IRC | 16:30 | |
| kfox1111 | ok. thanks. I think you folks have provided an alternate solution that will be even easier to implement, while providing the same level of flexability. So long as we can get https://blueprints.launchpad.net/neutron/+spec/dhcp-options-per-subnet implemented. | 16:31 |
| *** pal has quit IRC | 16:31 | |
| kfox1111 | I seem to remember a config file somewhere that you could override default options for dhcp in neutron... Remember where that is off hand? | 16:31 |
| kfox1111 | ah. I'm just blind. /etc/neutron/dnsmasq-neutron.conf | 16:33 |
| kfox1111 | perfect. I think this will work for what I need. :) | 16:35 |
| *** mlavalle has joined #openstack-neutron | 16:36 | |
| *** sudipto_ has quit IRC | 16:36 | |
| *** sudipto has quit IRC | 16:36 | |
| openstackgerrit | enikanorov proposed openstack/neutron: Avoid synchronizing session when deleting networkdhcpagentbinding https://review.openstack.org/171588 | 16:36 |
| *** johnbelamaric has joined #openstack-neutron | 16:37 | |
| *** ronis has joined #openstack-neutron | 16:39 | |
| *** dannywilson has quit IRC | 16:40 | |
| *** dannywilson has joined #openstack-neutron | 16:41 | |
| *** enikanorov_ has quit IRC | 16:42 | |
| *** Marga_ has quit IRC | 16:42 | |
| *** Marga_ has joined #openstack-neutron | 16:43 | |
| *** jpena is now known as jpena|away | 16:43 | |
| *** vthapar has quit IRC | 16:44 | |
| *** roeyc has quit IRC | 16:46 | |
| *** Marga_ has quit IRC | 16:46 | |
| *** chandankumar has quit IRC | 16:46 | |
| *** Marga_ has joined #openstack-neutron | 16:47 | |
| openstackgerrit | Merged openstack/neutron-fwaas: FWaaS: Reorganize unit test tree https://review.openstack.org/171602 | 16:48 |
| *** neeti has quit IRC | 16:48 | |
| *** Murali has quit IRC | 16:50 | |
| *** melwitt has joined #openstack-neutron | 16:51 | |
| *** julim has joined #openstack-neutron | 16:54 | |
| *** absubram has quit IRC | 16:54 | |
| *** Marga_ has quit IRC | 16:56 | |
| *** Marga_ has joined #openstack-neutron | 16:57 | |
| *** aepifanov has quit IRC | 16:59 | |
| *** moha_hunt has joined #openstack-neutron | 17:00 | |
| *** kawa2014 has quit IRC | 17:00 | |
| *** blues-man has quit IRC | 17:00 | |
| *** johnchalekson has quit IRC | 17:01 | |
| *** jckasper has left #openstack-neutron | 17:01 | |
| *** jckasper has joined #openstack-neutron | 17:01 | |
| *** clarkb has joined #openstack-neutron | 17:04 | |
| *** achanda has joined #openstack-neutron | 17:04 | |
| *** imcsk8 is now known as imcsk8|afk | 17:05 | |
| *** claudiub has joined #openstack-neutron | 17:05 | |
| *** _cjones_ has quit IRC | 17:06 | |
| *** _cjones_ has joined #openstack-neutron | 17:07 | |
| *** mspreitz has joined #openstack-neutron | 17:07 | |
| *** luqas has joined #openstack-neutron | 17:07 | |
| *** dannywilson has quit IRC | 17:07 | |
| mspreitz | If there are two shared external provider networks, what arrangement of routers and tenant networks would allow a tenant to have some floating IPs on each external network and all the tenant's Compute Instances to talk with each other? | 17:07 |
| armax | dane_leblanc, HenryG: ping | 17:08 |
| mspreitz | Using only one virtual NIC per Compute Instance | 17:08 |
| dane_leblanc | armax: Hi | 17:08 |
| *** pritesh has quit IRC | 17:08 | |
| *** tiswanso has quit IRC | 17:08 | |
| armax | mestery and I were trying to make sense of the most current plan of attack to address instability caused by: https://review.openstack.org/#/c/160622/ | 17:08 |
| armax | I see that HenryG +2 https://review.openstack.org/#/c/170690/ | 17:09 |
| *** tiswanso_ has joined #openstack-neutron | 17:09 | |
| *** nyechiel has quit IRC | 17:09 | |
| *** drjones has joined #openstack-neutron | 17:09 | |
| clarkb | I am trying to debug the neutron DVR multi node test and notice that when attaching a floating ip to a VM neutron adds the ip, arpings it (which fails), then removes the IP. This is problematic for a few reasons. Hard to debug because neutron has helpfully cleaned up the thing it decided was broken, and nova show reports that floating ip as being attached to the node without error | 17:09 |
| clarkb | is ^ a bug? | 17:09 |
| clarkb | or was this intentional and I need an education? | 17:09 |
| armax | but I am still trying to wrap my head around https://review.openstack.org/#/c/170690/ | 17:09 |
| *** _cjones_ has quit IRC | 17:10 | |
| armax | clarkb: ping | 17:10 |
| clarkb | armax: pong | 17:10 |
| armax | clarkb: hang on digesting what you wrote :) | 17:10 |
| *** yatin has joined #openstack-neutron | 17:10 | |
| dane_leblanc | armax: My feeling is that the commit https://review.openstack.org/#/c/160622/ exposed 2 problems. | 17:10 |
| *** annegentle has quit IRC | 17:10 | |
| armax | clarkb: do you see this error through manual testing? | 17:11 |
| dane_leblanc | armax: First, the DB deadlock timeout, which wasn't a problem with the #160622 per se, it's just that this commit exposed a problem in ML2, IMO | 17:11 |
| *** YorikSar has joined #openstack-neutron | 17:12 | |
| clarkb | armax: yes that is how I found it | 17:12 |
| openstackgerrit | Terry Wilson proposed openstack/neutron: Handle non-index lookups in native OVSDB backend https://review.openstack.org/167017 | 17:12 |
| *** luqas has quit IRC | 17:12 | |
| clarkb | I can paste the relevant logs from q-l3.log if it helps | 17:13 |
| dane_leblanc | armax: Second problem, the integrity error... I think this is something in the #160622 that may be unavoidable (port gets deleted by another thread right as we're updating it), or maybe there's a complicated way to avoid this, but it's no big deal to ignore the error. | 17:13 |
| *** tomoe_ has quit IRC | 17:13 | |
| armax | clarkb: that would be useful | 17:13 |
| *** rotbeard has quit IRC | 17:13 | |
| armax | clarkb: also, any configuration would be handy to see, for instance, how DVR has been configured | 17:14 |
| armax | dane_leblanc: is the integrity error shown only when using PG? | 17:14 |
| prometheanfire | ijw: I can't seem to set the mtu on network-create | 17:14 |
| HenryG | armax: sorry, was having lunch | 17:14 |
| ijw | prometheanfire: no, that's not a part of the patch functionality - we were asked not to do that | 17:15 |
| armax | clarkb: I was going to have a play with https://review.openstack.org/#/c/158525/ | 17:15 |
| ijw | You'll get the max mtu the system supports and it will be advertised to VMs | 17:15 |
| *** cbouch has quit IRC | 17:15 | |
| prometheanfire | ok | 17:15 |
| armax | clarkb: so I can help confirm the issue you’re seeing | 17:15 |
| dane_leblanc | armax: No, I'm seeing it on my VM using mysql | 17:15 |
| armax | clarkb: steps to repro would be good too | 17:15 |
| *** atuvenie has quit IRC | 17:16 | |
| *** yamahata has quit IRC | 17:16 | |
| armax | dane_leblanc: makes sense, so once we’re free from deadlock, we see the integrity error | 17:16 |
| *** tomoe_ has joined #openstack-neutron | 17:16 | |
| dane_leblanc | armax: Yes, 2nd layer of the onion. | 17:16 |
| armax | dane_leblanc: right, makes sense | 17:17 |
| clarkb | armax: steps to reproduce: use https://review.openstack.org/#/c/171452/4 to deploy multinode neutron with dvr, nova boot test, nova floating-ip-create public, nova floating-ip-associate test $addr | 17:17 |
| clarkb | getting the logs together now | 17:17 |
| armax | clarkb: gotcha, I’ll look into this | 17:17 |
| *** safchain has quit IRC | 17:18 | |
| *** yamahata has joined #openstack-neutron | 17:19 | |
| prometheanfire | ijw: so far it looks like the mtu is set everwhere but on the tap interface of the VM | 17:19 |
| prometheanfire | Sam-I-Am: ^ | 17:19 |
| clarkb | armax: http://paste.openstack.org/show/200579/ is the relevant bits of q-l3.log | 17:19 |
| *** tomoe_ has quit IRC | 17:19 | |
| prometheanfire | Sam-I-Am: that includes vxlan on nova-compute host | 17:19 |
| *** tomoe_ has joined #openstack-neutron | 17:19 | |
| armax | clarkb: looking | 17:20 |
| clarkb | armax: also fwiw I know that the network setup is broken in that test, so I was trying to reproduce manually to debug but ran into this | 17:20 |
| *** amotoki has quit IRC | 17:20 | |
| *** tomoe_ has quit IRC | 17:20 | |
| *** lukasa has joined #openstack-neutron | 17:21 | |
| *** harlowja_away is now known as harlowja | 17:21 | |
| *** ozialien has joined #openstack-neutron | 17:22 | |
| armax | clarkb: ok, let me see what’s going on, if dvr indeed cleans up what’s broken, we should reflect this all the way to nova | 17:22 |
| clarkb | armax: yes at the very least it should be "error blah blah X failed" | 17:22 |
| clarkb | and nova should show the floating ip as unassociated | 17:22 |
| *** gsagie_ has quit IRC | 17:23 | |
| clarkb | however, i am unsure how one is expected to debug this if neutron is going to do this | 17:23 |
| armax | clarkb: to be honest I am unaware of any cleanup logic in case of failure | 17:23 |
| armax | clarkb: there must be something introduced recently | 17:23 |
| armax | clarkb: as you obviosuly point out, if neutron pulls the rug from one’s feet it’s very difficult to see what’s going | 17:24 |
| clarkb | armax: not to completely change the subject, but now that I understand some of this stuff a bit better https://review.openstack.org/#/c/171714/ is sort of where I think we should go for the test environment overlay setup | 17:25 |
| clarkb | (not currently working but thats the general shape of the thing) | 17:25 |
| *** claudiub has quit IRC | 17:26 | |
| *** Sukhdev has quit IRC | 17:26 | |
| armax | clarkb: I thought you didn’t want to special case the setup for neutron? | 17:27 |
| armax | clarkb: what am I missing? | 17:27 |
| clarkb | armax: I don't, nova-net is using that same function in that change too | 17:27 |
| *** vivek-ebay has joined #openstack-neutron | 17:27 | |
| openstackgerrit | Edgar Magana proposed openstack/neutron: Add missing config parameters in neutron.conf https://review.openstack.org/171059 | 17:27 |
| * armax looks more closely | 17:27 | |
| *** atuvenie has joined #openstack-neutron | 17:27 | |
| armax | clarkb: yup, sorry | 17:27 |
| clarkb | armax: so basically nova net gets an ovs_gre_bridge for the "flat" network and for the "public" network and in neutron case we only do the br-ex network | 17:27 |
| *** armax has quit IRC | 17:29 | |
| *** armax has joined #openstack-neutron | 17:29 | |
| armax | clarkb: sorry my laptop decided to die | 17:30 |
| clarkb | armax: but my ip link manipulation is currently broken, so need to figure that out | 17:30 |
| clarkb | armax: np :) | 17:30 |
| armax | clarkb: did I miss anything? | 17:30 |
| clarkb | armax: nope | 17:30 |
| armax | ok | 17:30 |
| clarkb | but yes I need to delete the old gre bridge function once I get the ovs stuff working | 17:31 |
| clarkb | that should hopefully make my change a bit more clear | 17:31 |
| armax | I’ll poke at this a bit to see what’s going on…I really want to get this multi-node going | 17:31 |
| HenryG | marun: ping | 17:34 |
| marun | HenryG: pong | 17:34 |
| armax | dane_leblanc: is the rate of the IntegrityError failure the same of the deadlock? | 17:34 |
| HenryG | marun: Is this a known intermittent functional test failure, or something new? http://logs.openstack.org/90/170690/7/check/check-neutron-dsvm-functional/52befd9/console.html#_2015-04-08_15_42_42_536 | 17:34 |
| armax | dane_leblanc: in other words ahve we traded one trace for another>? | 17:35 |
| *** yatin has quit IRC | 17:35 | |
| marun | HenryG: we've been seeing that more often | 17:35 |
| marun | HenryG: We bumped the timeout from 60s to 90s, which helped | 17:35 |
| marun | HenryG: If it's still occurring at 90s, we may need to bump that limit in tox.ini | 17:35 |
| HenryG | marun: So it eventually respawn? | 17:36 |
| dane_leblanc | armax: It's hard to tell, but I think the rate of occurance of the integrity error is lower, maybe half the occurance rate? | 17:36 |
| marun | HenryG: It should - none of the code in question has changed | 17:36 |
| *** drjones has quit IRC | 17:36 | |
| marun | HenryG: My guess is we are seeing interaction at the ovs level due to the new ovs functional tests | 17:36 |
| marun | HenryG: interaction -> contention | 17:36 |
| marun | HenryG: Without separate ovs sandboxes per worker, I think we're stuck with bumping the timeout to something that works consistently. | 17:37 |
| armax | dane_leblanc: ok | 17:37 |
| HenryG | marun: ok | 17:37 |
| *** suresh12 has joined #openstack-neutron | 17:38 | |
| *** _cjones_ has joined #openstack-neutron | 17:39 | |
| *** yamahata has quit IRC | 17:39 | |
| *** achanda has quit IRC | 17:40 | |
| openstackgerrit | Edgar Magana proposed openstack/neutron: Add missing config parameters in neutron.conf https://review.openstack.org/171059 | 17:40 |
| HenryG | armax: oops sorry I rechecked on top of your vote | 17:41 |
| armax | wtf | 17:41 |
| armax | sorry | 17:41 |
| armax | that should be ok | 17:41 |
| armax | I see these errors on the PG job: http://logs.openstack.org/68/170968/3/check/check-tempest-dsvm-neutron-pg-full/f1ebabb/logs/screen-q-svc.txt.gz?level=TRACE | 17:41 |
| armax | dane_leblanc, HenryG: I wonder if we have a third layer of exception triggered now | 17:42 |
| *** tomoe_ has joined #openstack-neutron | 17:42 | |
| dane_leblanc | armax: The newest error says that a call to rollback is necessary after there's an exception during the flush (and flush was just added)... | 17:43 |
| dane_leblanc | armax: I'm guessing this is a PG thing? | 17:43 |
| armax | dane_leblanc: that’s not how I interpreted it | 17:43 |
| dane_leblanc | armax: In my testing with mysql, the original integrity error was caught without issue. | 17:44 |
| armax | it looks like we’re not using the transaction properly | 17:44 |
| armax | let me inspect the code a bit closer | 17:44 |
| dane_leblanc | armax: Sure. Always open to suggestions/help! | 17:45 |
| armax | dane_leblanc: everything comes at a cost, though! | 17:45 |
| armax | dane_leblanc: I think we’re prone to that issue with mysql too | 17:47 |
| armax | dane_leblanc: your fix is no good | 17:48 |
| armax | dane_leblanc: I think we need to look at it a bit closer | 17:48 |
| armax | dane_leblanc: because you’re catching an error in a transaction that involve multiple DB statements | 17:49 |
| armax | dane_leblanc: as many as the ports being yielded by the port_qry | 17:49 |
| armax | we can’t reause a transaction that’s in error | 17:49 |
| armax | that’s what the PG job is barfing about | 17:49 |
| *** tomoe_ has quit IRC | 17:50 | |
| armax | and the mysql one would be just as a prone | 17:50 |
| prometheanfire | Sam-I-Am: I can ping, but not ssh, I lowered the mtu of the dhcp namespace's interface, and still can't ssh | 17:50 |
| prometheanfire | Sam-I-Am: mtu advertized in dnsmasq is 1450 still (wanted to change that last) | 17:50 |
| *** Guest36 has joined #openstack-neutron | 17:52 | |
| *** achanda has joined #openstack-neutron | 17:53 | |
| *** dboik_ has joined #openstack-neutron | 17:53 | |
| dane_leblanc | armax: Is there a way to catch an error in a transaction and continue, without trashing the entire transaction? | 17:53 |
| armax | dane_leblanc: the use of savepoints would do | 17:53 |
| armax | dane_leblanc: but that won’t be liked by the UT | 17:54 |
| armax | dane_leblanc: I am afraid, our last resort, as much as I hate | 17:54 |
| armax | is to revert your change | 17:54 |
| armax | dane_leblanc: we can’t afford another minute of gate instability | 17:54 |
| armax | dane_leblanc: and there’s more to work on | 17:54 |
| dane_leblanc | armax: Yep | 17:54 |
| armax | we tried | 17:54 |
| armax | but we’re not ready yet | 17:55 |
| armax | let’s restore peace and work on a better solution in isolation | 17:55 |
| armax | dane_leblanc, HenryG, mestery: seems fair? ^^^ | 17:55 |
| *** tomoe_ has joined #openstack-neutron | 17:55 | |
| HenryG | armax: makes sense | 17:56 |
| *** dboik has quit IRC | 17:56 | |
| HenryG | armax: can this stay? https://review.openstack.org/170690 | 17:56 |
| armax | HenryG: I think so | 17:56 |
| dane_leblanc | armax: Sure. Could use some help on narrowing in on the ultimate fix. | 17:56 |
| prometheanfire | ijw: ping | 17:57 |
| armax | how bad if we ship without a fix for bug #1427474 | 17:57 |
| openstack | bug 1427474 in neutron "IPv6 SLAAC subnet create should update ports on net" [Medium,Fix committed] https://launchpad.net/bugs/1427474 - Assigned to Dane LeBlanc (leblancd) | 17:57 |
| Sam-I-Am | prometheanfire: mmmkay | 17:57 |
| *** rangansw has joined #openstack-neutron | 17:58 | |
| *** atuvenie has quit IRC | 17:58 | |
| openstackgerrit | Armando Migliaccio proposed openstack/neutron: Revert "IPv6 SLAAC subnet create should update ports on net" https://review.openstack.org/171761 | 17:58 |
| *** bjornar_ has joined #openstack-neutron | 17:58 | |
| armax | HenryG: ^^ | 17:58 |
| mestery | armax: ++ | 17:58 |
| dane_leblanc | armax: Without this fix, if anyone adds a SLAAC subnet AFTER VMs are instantiated, the VMs won't be updated with SLAAC | 17:58 |
| armax | mestery: ^^ | 17:58 |
| armax | dane_leblanc: is there a workaround? | 17:59 |
| *** yamahata has joined #openstack-neutron | 17:59 | |
| mestery | dane_leblanc: Can we release note that? | 17:59 |
| *** tsg_ has joined #openstack-neutron | 17:59 | |
| armax | dane_leblanc: if we’d only fix the damn test, we could release your fix | 17:59 |
| dane_leblanc | mestery: Sure, how do I create a release note. | 18:00 |
| mestery | dane_leblanc: Let me find the Kilo release note section | 18:00 |
| armax | because I still believe that in prod the circumstances where the deadlock is triggered are pretty exceptional | 18:00 |
| armax | dane_leblanc: do you guys know the test author? | 18:00 |
| armax | or is anyone interested in cleaning that one up? | 18:00 |
| *** annegentle has joined #openstack-neutron | 18:01 | |
| dane_leblanc | armax: Yes, but he's in another group now (I think it was Sergey?) | 18:01 |
| armax | that test_dhcpv6 sends shivers up my spine | 18:01 |
| dane_leblanc | armax: But it is just evil enough to catch some real issues. | 18:02 |
| armax | dane_leblanc: agreed | 18:02 |
| armax | there are many ways we can tip Neutron over the edge | 18:02 |
| armax | that doesn’t mean we need to have them in the gate | 18:02 |
| armax | and make everyone miserable | 18:03 |
| dane_leblanc | armax: Any examples on savepoints you can point me to? | 18:03 |
| armax | none in the code | 18:03 |
| armax | because UT won’t like it | 18:03 |
| *** Aish has joined #openstack-neutron | 18:04 | |
| *** rushil_ has joined #openstack-neutron | 18:04 | |
| prometheanfire | Sam-I-Am: traceroute with 1000b packets fails too, which is kinda odd | 18:05 |
| Sam-I-Am | prometheanfire: eeeoo | 18:05 |
| prometheanfire | pings work though | 18:05 |
| prometheanfire | so... od | 18:05 |
| *** seizadi has joined #openstack-neutron | 18:06 | |
| *** seizadi1 has joined #openstack-neutron | 18:06 | |
| *** seizadi has quit IRC | 18:06 | |
| *** seizadi1 has quit IRC | 18:07 | |
| *** rushil has quit IRC | 18:07 | |
| armax | dane_leblanc, HenryG, mestery: | 18:08 |
| armax | I updated reports https://bugs.launchpad.net/neutron/+bug/1441382 https://bugs.launchpad.net/neutron/+bug/1440183 and https://bugs.launchpad.net/neutron/+bug/1427474 | 18:08 |
| openstack | Launchpad bug 1441382 in neutron "IPv6 SLAAC subnet Tempest tests fail due to IntegrityError " [High,Incomplete] - Assigned to Dane LeBlanc (leblancd) | 18:08 |
| openstack | Launchpad bug 1440183 in neutron "DBDeadlock on subnet allocation" [Critical,Incomplete] - Assigned to Dane LeBlanc (leblancd) | 18:08 |
| openstack | Launchpad bug 1427474 in neutron "IPv6 SLAAC subnet create should update ports on net" [Medium,Confirmed] - Assigned to Dane LeBlanc (leblancd) | 18:08 |
| *** seizadi has joined #openstack-neutron | 18:08 | |
| armax | dane_leblanc: please resubmit your fix and let’s try to kill the failure modes in the new patch | 18:08 |
| mestery | armax: thanks! | 18:09 |
| dane_leblanc | armax: | 18:09 |
| armax | dane_leblanc: I mean the fix being reverted | 18:09 |
| armax | dane_leblanc: with any additional logic to deal with the exposed failure mdoes | 18:09 |
| *** sam_stone has joined #openstack-neutron | 18:09 | |
| HenryG | armax: thanks for all your help here | 18:09 |
| armax | HenryG: sure, I wish we could have nailed this one | 18:10 |
| dane_leblanc | armax: Right, will do! If there's another way to address handling an exception in the middle of a multi entry DB transaction, please let me know. | 18:10 |
| armax | dane_leblanc: I think we’d need to get rid of the subtransaction | 18:11 |
| armax | dane_leblanc: but that might lead to other funky failure modes :) | 18:11 |
| armax | dane_leblanc: that’s why I didn’t suggest it in the review | 18:11 |
| dane_leblanc | armax: The subtransaction? I don't have that in the latest patch. | 18:12 |
| armax | it’s in the caller methods | 18:12 |
| armax | create_subnet_db and create_subnet_implicit pool | 18:12 |
| dane_leblanc | armax: Ah, was just going to ask that. | 18:13 |
| *** tomoe_ has quit IRC | 18:15 | |
| *** dboik_ has quit IRC | 18:16 | |
| HenryG | armax: you know we do have access to zzzeek. Although it could be embarrassing when he looks at neutron code, we might learn a lot | 18:16 |
| zzzeek | HenryG: there’s a particular spot in neutron I want to clean up but havent had time to get into it more deeply | 18:16 |
| HenryG | zzzeek: do tell | 18:17 |
| armax | hi zzzeek | 18:17 |
| zzzeek | armax: hey | 18:17 |
| zzzeek | HenryG: well its not a place that i think people find to be that important but its a general pattern i want to clean up across projects, let me find it | 18:18 |
| *** chandankumar has joined #openstack-neutron | 18:18 | |
| *** dannywilson has joined #openstack-neutron | 18:19 | |
| zzzeek | HenryG: this thing: https://github.com/openstack/neutron/blob/master/neutron/plugins/ml2/plugin.py#L486 | 18:19 |
| zzzeek | HenryG: this is in a few places. the model_query() for neutron is broken up into these three awkward phases, and seveal of these pliugins put an unnecessary and expensive OUTER JOIN on all queries | 18:19 |
| *** ankit_ag has quit IRC | 18:19 | |
| zzzeek | HenryG: this should be an INNER JOIN and only when the filter_hook is actually in use | 18:20 |
| zzzeek | now its hard for me to change this b.c. everyone will be like, it works great and nobody uses that thing so who cares | 18:20 |
| zzzeek | HenryG: but i really want to fix up how we build queries to be cleaner, using newer techniques | 18:21 |
| HenryG | zzzeek: I will file a wishlist bug and work with you to get it done | 18:22 |
| *** SumitNaiksatam has joined #openstack-neutron | 18:22 | |
| zzzeek | HenryG: there’s a quick cahnge we can make right there that will probably corect the outerjoin, we can do query.join() right in the _ml2_port_result_filter_hook for now | 18:23 |
| amuller | zzzeek: Is the performance impact measurable? | 18:23 |
| HenryG | zzzeek: Good to know, but the release is being cut so not today :) | 18:24 |
| *** yeungp has joined #openstack-neutron | 18:24 | |
| *** shashankhegde has joined #openstack-neutron | 18:24 | |
| zzzeek | amuller: ive no idea because i haven’t studied how many rows that area is working on. but my motiviation is more like, “less code that is also simpler to read” | 18:24 |
| *** aix has quit IRC | 18:24 | |
| zzzeek | amuller: I want to come up with a simpler model_query() across all projects. neutron’s model_query() seems like the most complex | 18:25 |
| amuller | We're so 'special' :) | 18:25 |
| zzzeek | amuller: not really :) i wanted to clean up transaction logic, and nova wins the prize on that one :) | 18:25 |
| amuller | zzzeek: The number one issue from my perspective is how we manage DB transactions and notifications | 18:28 |
| *** annegentle has quit IRC | 18:28 | |
| *** boris-42 has quit IRC | 18:28 | |
| zzzeek | amuller: neutron seemed to have a reasonable approach to transaction scope, seems like it starts it up at the beginning of a request and keeps it open til the end…e.g isn’t calling get_session() all over the place | 18:29 |
| zzzeek | amuller: if i recall correctly | 18:29 |
| *** GLaupre has joined #openstack-neutron | 18:29 | |
| amuller | zzzeek: Where we try quite hard during reviews to ensure that notifications via API or RPC are sent after transactions are commited, but there's cases where those slip in, cause issues at the gate, then we revert... | 18:29 |
| zzzeek | amuller: I’d assume you add notifcations into a callback framework of some kind that is linked to transaction scope completing | 18:30 |
| *** annegentle has joined #openstack-neutron | 18:30 | |
| amuller | zzzeek: yeah, we don't :) | 18:30 |
| zzzeek | amuller: well that woudl be how to do tha t:) | 18:30 |
| *** amrith has left #openstack-neutron | 18:31 | |
| *** vivek-ebay has quit IRC | 18:31 | |
| amuller | zzzeek: it was what I was about to bring up... have some sort of framework where you can queue up notifications and have them automatically send out when the relevant transaction is committed | 18:31 |
| amuller | zzzeek: but then... you get into the awful realm of distributed state where it looks like none of us really know or have the energy to deal with | 18:31 |
| amuller | zzzeek: what if you commit, send out the notification, and that fails | 18:32 |
| zzzeek | amuller: so that is what a system like zope.transaction does, maybe look at their architecture and build something much simpler than that | 18:32 |
| HenryG | zzzeek: there are still a couple of places where get_session() is called. They are on the radar for cleanup. | 18:32 |
| amuller | zzzeek: we have a ton of plugins that do use the neutron DB and send out http calls to the vendor API server | 18:33 |
| zzzeek | amuller: yeah they should all use some kind of framework so that the plugin itself is constructed declaratively | 18:33 |
| amuller | zzzeek: can you expand? or link to some document? | 18:34 |
| zzzeek | amuller: well it means, instead of writing imperative code: session.do_thing(); session.commit(); web_service.send_the_thing(); you have functions that define “db work” and then functions that define things that happen after the DB, and they are structured either on top of a template class or even as data structures of callbacks | 18:35 |
| zzzeek | amuller: so that the plugin doesnt decide when to send out hte messages, the plugin container knows how to orchestrate the steps | 18:35 |
| openstackgerrit | Aaron Rosen proposed openstack/neutron: _create_subnet_from_implicit_pool assumes external network extension https://review.openstack.org/171770 | 18:35 |
| *** chandankumar has quit IRC | 18:36 | |
| amuller | zzzeek: yeah gotcha | 18:36 |
| GLaupre | Hello all, can someone explain me something concering ML2 + MD openvswitch for VLAN? | 18:38 |
| GLaupre | I don't want to use tunneling, only vlans, why we need 2 bridges anyway? (br-int and br-<physical interface>) : http://docs.openstack.org/admin-guide-cloud/content/figures/14/a/a/common/figures/under-the-hood-scenario-1-ovs-compute.png | 18:40 |
| GLaupre | Can't we combine the rules for translating global/local tag in the same bridge instead of using 2? | 18:40 |
| *** dboik has joined #openstack-neutron | 18:41 | |
| *** dboik has quit IRC | 18:41 | |
| *** marrusl has joined #openstack-neutron | 18:41 | |
| *** dboik has joined #openstack-neutron | 18:42 | |
| *** dave-mccowan has quit IRC | 18:47 | |
| lifeless | GLaupre: one is logical one physical, and neutron then deals solely in logical definitions | 18:51 |
| lifeless | GLaupre: if there was one we'd need a more complex language to describe the physical interconnect on a per-agent basis to Neutron - way more complex | 18:51 |
| *** annegentle has quit IRC | 18:53 | |
| *** pritesh has joined #openstack-neutron | 18:53 | |
| GLaupre | lifeless: can you explain me a bit more the dif between the logical (br-int) and physical (br-eth), in what their job is different? | 18:53 |
| *** sbalukoff has joined #openstack-neutron | 18:54 | |
| *** vivek-ebay has joined #openstack-neutron | 18:55 | |
| *** vivek-ebay has quit IRC | 18:57 | |
| *** matrohon has joined #openstack-neutron | 18:58 | |
| *** dboik_ has joined #openstack-neutron | 19:05 | |
| *** dboik has quit IRC | 19:05 | |
| *** tidwellr has joined #openstack-neutron | 19:08 | |
| *** atuvenie has joined #openstack-neutron | 19:09 | |
| *** luqas has joined #openstack-neutron | 19:09 | |
| *** ociuhandu_ has quit IRC | 19:10 | |
| *** imcsk8|afk is now known as imcsk8 | 19:10 | |
| *** Marga_ has quit IRC | 19:12 | |
| *** luqas has quit IRC | 19:14 | |
| *** annegentle has joined #openstack-neutron | 19:15 | |
| *** tomoe_ has joined #openstack-neutron | 19:15 | |
| *** igordcard_ has joined #openstack-neutron | 19:18 | |
| *** suresh12 has quit IRC | 19:19 | |
| *** shashankhegde has quit IRC | 19:19 | |
| *** tomoe_ has quit IRC | 19:20 | |
| *** ivar-lazzaro has quit IRC | 19:21 | |
| *** aepifanov has joined #openstack-neutron | 19:22 | |
| *** ivar-lazzaro has joined #openstack-neutron | 19:22 | |
| *** matrohon has quit IRC | 19:29 | |
| *** jprovazn has quit IRC | 19:34 | |
| *** ivar-lazzaro has quit IRC | 19:34 | |
| *** achanda has quit IRC | 19:37 | |
| *** thedodd has quit IRC | 19:37 | |
| *** Sukhdev has joined #openstack-neutron | 19:41 | |
| *** absubram has joined #openstack-neutron | 19:44 | |
| *** achanda has joined #openstack-neutron | 19:46 | |
| *** wznoinsk has quit IRC | 19:47 | |
| *** wznoinsk has joined #openstack-neutron | 19:49 | |
| *** dkehn has quit IRC | 19:50 | |
| *** _cjones_ has quit IRC | 19:50 | |
| *** enikanorov has joined #openstack-neutron | 19:51 | |
| *** ZZelle_ has joined #openstack-neutron | 19:51 | |
| *** dkehn has joined #openstack-neutron | 19:51 | |
| *** yamamoto has joined #openstack-neutron | 19:52 | |
| *** nplanel_ has quit IRC | 19:52 | |
| enikanorov | dougwig: hi, could you please review https://review.openstack.org/#/c/171567/ | 19:53 |
| *** xgerman has quit IRC | 19:53 | |
| *** xgerman has joined #openstack-neutron | 19:54 | |
| *** shashankhegde has joined #openstack-neutron | 19:54 | |
| openstackgerrit | Armando Migliaccio proposed openstack/neutron: Update L3 Agent Scheduler API tests https://review.openstack.org/171794 | 19:54 |
| *** ijw has quit IRC | 19:56 | |
| *** jpena|away is now known as jpena | 19:56 | |
| *** yamamoto has quit IRC | 19:57 | |
| dougwig | enikanorov: done | 19:57 |
| *** jorgem has joined #openstack-neutron | 19:58 | |
| *** kbringard has quit IRC | 19:58 | |
| *** thedodd has joined #openstack-neutron | 19:59 | |
| *** ozialien has quit IRC | 19:59 | |
| *** zzzeek has quit IRC | 20:04 | |
| *** zzzeek_ has joined #openstack-neutron | 20:04 | |
| *** matrohon has joined #openstack-neutron | 20:04 | |
| *** _cjones_ has joined #openstack-neutron | 20:04 | |
| *** achanda has quit IRC | 20:05 | |
| *** pritesh has quit IRC | 20:05 | |
| *** ivar-lazzaro has joined #openstack-neutron | 20:06 | |
| clarkb | armax https://jenkins04.openstack.org/job/check-tempest-dsvm-neutron-aiopcpu-full/22/console failed even with the ovs setup of br-ex | 20:09 |
| armax | clarkb: looking | 20:10 |
| *** bradjones_ has joined #openstack-neutron | 20:11 | |
| *** Marga_ has joined #openstack-neutron | 20:11 | |
| *** dkehn has quit IRC | 20:12 | |
| *** bradjones has quit IRC | 20:13 | |
| *** bradjones_ is now known as bradjones | 20:13 | |
| *** bradjones has joined #openstack-neutron | 20:13 | |
| *** ivar-lazzaro has quit IRC | 20:14 | |
| *** dave-mccowan has joined #openstack-neutron | 20:14 | |
| *** kbringard has joined #openstack-neutron | 20:14 | |
| *** kbringard1 has joined #openstack-neutron | 20:14 | |
| *** ivar-lazzaro has joined #openstack-neutron | 20:14 | |
| *** ivar-lazzaro has quit IRC | 20:15 | |
| *** dttocs has joined #openstack-neutron | 20:16 | |
| *** vivek-ebay has joined #openstack-neutron | 20:16 | |
| *** ivar-lazzaro has joined #openstack-neutron | 20:16 | |
| *** annegentle has quit IRC | 20:17 | |
| *** kbringard has quit IRC | 20:18 | |
| *** prad has quit IRC | 20:19 | |
| *** tomoe_ has joined #openstack-neutron | 20:21 | |
| *** annegentle has joined #openstack-neutron | 20:23 | |
| *** tomoe_ has quit IRC | 20:25 | |
| *** tsg_ has quit IRC | 20:26 | |
| *** smccully has quit IRC | 20:26 | |
| openstackgerrit | Merged openstack/neutron: Moving VLAN Transparency support from core to extension https://review.openstack.org/169569 | 20:30 |
| *** dkehn has joined #openstack-neutron | 20:30 | |
| *** crc32 has joined #openstack-neutron | 20:30 | |
| *** prad has joined #openstack-neutron | 20:30 | |
| openstackgerrit | Kevin Benton proposed openstack/neutron: Add simple ARP spoofing protection https://review.openstack.org/171003 | 20:31 |
| *** mwagner_lap has quit IRC | 20:32 | |
| kevinbenton | carl_baldwin: did a quick re-spin to switch to items ^^ | 20:32 |
| *** pritesh has joined #openstack-neutron | 20:33 | |
| carl_baldwin | kevinbenton, ack | 20:34 |
| *** Guest36 has quit IRC | 20:38 | |
| *** annegentle has quit IRC | 20:38 | |
| *** annegentle has joined #openstack-neutron | 20:39 | |
| *** mmdurrant has joined #openstack-neutron | 20:39 | |
| *** achanda has joined #openstack-neutron | 20:39 | |
| openstackgerrit | Merged openstack/neutron: Revert "IPv6 SLAAC subnet create should update ports on net" https://review.openstack.org/171761 | 20:40 |
| mmdurrant | There's either a pip package or apt package for neutron-lbaas-plugin-agent and I cannot remember the name of the damned thing for the life of me. | 20:40 |
| *** tsufiev has left #openstack-neutron | 20:42 | |
| pritesh | mestery: Thanks. | 20:42 |
| mestery | pritesh: yw :) | 20:42 |
| *** dttocs_ has joined #openstack-neutron | 20:42 | |
| *** apuimedo has joined #openstack-neutron | 20:43 | |
| *** annegentle has quit IRC | 20:45 | |
| *** dttocs has quit IRC | 20:47 | |
| *** sbfox has joined #openstack-neutron | 20:47 | |
| *** jpena has quit IRC | 20:48 | |
| *** vivek-ebay has quit IRC | 20:51 | |
| *** pritesh has quit IRC | 20:51 | |
| *** Marga_ has quit IRC | 20:51 | |
| *** Marga_ has joined #openstack-neutron | 20:53 | |
| *** Marga_ has quit IRC | 20:54 | |
| *** Marga_ has joined #openstack-neutron | 20:54 | |
| *** raildo is now known as raildo|away | 20:54 | |
| pc_m | mestery: hi | 20:55 |
| *** dttocs_ has quit IRC | 20:55 | |
| mestery | armax HenryG: I see you guys merged this one, but the bug isn't targeting RC1: https://bugs.launchpad.net/neutron/+bug/1440183 | 20:56 |
| openstack | Launchpad bug 1440183 in neutron "DBDeadlock on subnet allocation" [Critical,Incomplete] - Assigned to Dane LeBlanc (leblancd) | 20:56 |
| mestery | https://review.openstack.org/#/c/170690/ | 20:56 |
| mestery | armax HenryG: ^^^ | 20:56 |
| HenryG | mestery: sorry for the miscommunication | 20:57 |
| armax | mestery: the question is this | 20:57 |
| mestery | HenryG: Does that resolve the bug then? | 20:57 |
| mestery | Is it partial? | 20:57 |
| mestery | I've added it to the RC, and shed many tears in the process | 20:57 |
| mestery | :( | 20:57 |
| armax | we should have edited the commit message to remove the closes-bug thingy | 20:57 |
| mestery | OK | 20:57 |
| mestery | armax: So, leave the bug open for Liberty-1? | 20:57 |
| mestery | I'm fine with that | 20:57 |
| *** absubram has quit IRC | 20:57 | |
| armax | ultimately the revert patch killed that bug | 20:58 |
| openstackgerrit | enikanorov proposed openstack/neutron: Avoid synchronizing session when deleting networkdhcpagentbinding https://review.openstack.org/171588 | 20:58 |
| mestery | Right | 20:58 |
| armax | see https://bugs.launchpad.net/neutron/+bug/1440183 last comment | 20:58 |
| openstack | Launchpad bug 1440183 in neutron "DBDeadlock on subnet allocation" [Critical,Confirmed] - Assigned to Dane LeBlanc (leblancd) | 20:58 |
| mestery | Now that will be a part of Kilo, which is ok | 20:58 |
| armax | the one that Henry and I merged | 20:58 |
| mestery | armax: Done and done | 20:58 |
| armax | we could pull it out of the gate queue, but that really should not hurt | 20:58 |
| armax | the only nuisance is hte commit message which is misleading | 20:58 |
| mlavalle | carl_baldwin: so, is tomorrow's meeting cancelled? | 20:59 |
| openstackgerrit | enikanorov proposed openstack/neutron: Add logging to dangling port to ml2 delete_subnet https://review.openstack.org/171662 | 20:59 |
| carl_baldwin | mlavalle, as far as I know, yes. | 20:59 |
| mlavalle | carl_baldwin: cool, see you there next week :-) | 21:00 |
| mestery | armax: Lets leave it in, it's fine | 21:00 |
| *** dkehn has quit IRC | 21:00 | |
| armax | ok | 21:00 |
| *** sbfox has quit IRC | 21:00 | |
| *** annegentle has joined #openstack-neutron | 21:00 | |
| pc_m | Anyone: Is there a know issue in neutron with import paths still? I happen to be playing in VPN, with an import of neutron.tests.functional.agent.test_l3_agent, and it is importing neutron.tests.unit.test_l3_agent (which is wrong). Didn't know if someone is already on this? | 21:01 |
| mestery | armax: So, we have emagana's config fix left, and then we get to rattle on kevinbenton's arp fix and if we should merge it all or just the config option | 21:01 |
| mestery | armax: Then Kilo is cooked | 21:01 |
| armax | ok | 21:01 |
| emagana | (y) | 21:01 |
| HenryG | w00t! | 21:02 |
| * amuller tries to find some critical bug that has to be fixed for Kilo just to watch mestery's response | 21:04 | |
| * mestery shakes his fist at amuller | 21:04 | |
| mestery | amuller: For each additional bug you find, you owe me one bottle of Bombay Saphire Gin ;) | 21:04 |
| enikanorov | i have a few, but not critical enough to spoil the party ) | 21:04 |
| mestery | enikanorov: :) | 21:04 |
| openstackgerrit | Sanjeev Rampal proposed openstack/neutron: Improves the description string for the config parameter metadata_workers. This will get extracted into the documentation automatically https://review.openstack.org/171026 | 21:05 |
| mestery | Awesome work folks! Seriously, a great effort! | 21:05 |
| pc_m | mestery: ping | 21:06 |
| mestery | pc_m: pong | 21:07 |
| pc_m | mestery: See my question a few mins ago? | 21:07 |
| mestery | pc_m: No, but reading now | 21:07 |
| pc_m | mestery: About imports... | 21:07 |
| mestery | pc_m: Try removing your .tox/py27 directory perhaps? | 21:08 |
| HenryG | pc_m: let me check | 21:08 |
| clarkb | armax: I see your test of swami's change without my edits, will be good to see how that does | 21:08 |
| armax | clarkb: yea..I am just fishing for more data points | 21:09 |
| pc_m | mestery: Just looking at the code it is not right in my repo that I did a pull on. | 21:09 |
| armax | in the meantime I am also creating my own env | 21:09 |
| armax | so that I can make more sense as on what’s going on | 21:09 |
| *** annegentle has quit IRC | 21:09 | |
| emagana | Folks: should be very straight forward to review and merge the config change: https://review.openstack.org/#/c/171059/ | 21:09 |
| clarkb | armax: hrm looks like it failed bceause ovs-vsctl isn't installed. that was one of the things I fixed in my change | 21:09 |
| emagana | I am around to address any suggestion and comments right away! | 21:09 |
| mestery | HenryG: See comments from emagana above | 21:09 |
| armax | clarkb: which one? | 21:09 |
| mestery | HenryG marun: Config change review! ^^^^ :) | 21:09 |
| HenryG | pc_m: you are right | 21:10 |
| *** annegentle has joined #openstack-neutron | 21:10 | |
| clarkb | armax: https://review.openstack.org/#/c/171714/ that one and its parent fix several problems in swami's original change | 21:10 |
| pc_m | HenryG: I don't know why it isn't breaking the functional test. | 21:10 |
| clarkb | armax: the only reason it tested out good before is the DVR flag wasn't set so it wasn't self testing | 21:10 |
| *** lukasa has quit IRC | 21:10 | |
| armax | clarkb: right | 21:10 |
| *** luqas has joined #openstack-neutron | 21:10 | |
| HenryG | pc_m: not sure, investigating | 21:11 |
| clarkb | armax: I think 171714 is actually pretty close now | 21:11 |
| armax | clarkb: I thought that ovs was going to be installed automatically by Devstack | 21:11 |
| *** jckasper has quit IRC | 21:11 | |
| clarkb | armax: this happens before devstack runs | 21:11 |
| armax | clarkb: but if we install it explictely I don’t think it hurts | 21:11 |
| clarkb | armax: this is why I really wanted devstack to do this for us but it appears that it does not | 21:12 |
| clarkb | armax: sdague says that the new lib/neutron in devstack will likely do this for us properly but the existing lib/neutron-legacy does not | 21:12 |
| armax | clarkb: it should | 21:12 |
| armax | lib/neutron_plugins/ovs_base | 21:12 |
| armax | in lib/neutron_plugins/ovs_base | 21:12 |
| pc_m | HenryG: So bottom line is no one is addressing it. Shall I fix it? | 21:12 |
| *** pradipta has quit IRC | 21:12 | |
| mestery | pc_m: Yes please :) | 21:12 |
| clarkb | armax: yes that all happens too late and you can't tell it attach $interface to br-ex | 21:12 |
| mestery | pc_m: If it's an issue, shoot a patch and we'll land it soon | 21:12 |
| marun | mestery: hand coded? :/ | 21:13 |
| pc_m | OK. will do now... | 21:13 |
| armax | clarkb: oh, right gotcha | 21:13 |
| clarkb | armax: it is expected that I the user setup br-ex for neutron but not br-int | 21:13 |
| HenryG | pc_m: sure, but we should understand why it doesn't break the functional tests | 21:13 |
| mestery | marun: Yes! I expect emagana to be able to propose the tool to generate for Kilo | 21:13 |
| clarkb | its all very confusing for the noob and should be corrected | 21:13 |
| mestery | marun: But we're at the stage where we just need it fixed for Kilo now. Make sense? | 21:13 |
| armax | so really my change https://review.openstack.org/#/c/171809/ is too much behind your troubleshooting curve | 21:13 |
| pc_m | HenryG: YEs. Can you look at that, and I'll proceed on the fix. | 21:13 |
| clarkb | armax: I think so | 21:13 |
| marun | mestery: ok | 21:13 |
| mestery | marun: It's not ideal, I know, but we just need it done | 21:13 |
| mestery | For Kilo | 21:13 |
| mestery | FOR KILO! | 21:13 |
| mestery | Go Kilo! | 21:13 |
| mestery | :) | 21:14 |
| emagana | marun: Yes! I tried to implement the same script that nova team uses we I found a few issues that will require change in the neutron code | 21:14 |
| marun | mestery: I just don't know how to review this kind of thing. | 21:14 |
| emagana | mostly the way we name the config parameters | 21:14 |
| marun | mestery: I'm not as diligent as ihar - I can't do much more than rubber-stamp. | 21:14 |
| clarkb | armax: a quick scan of my two changes should show you what I have changed, its not a whole lot. Basically had to fix a bug enabling q-l3 on the subnode, had to add q-fwaas and q-meta to the subnode, and finally the bugs in the ovs bridge setup | 21:14 |
| armax | clarkb: I saw | 21:14 |
| mestery | marun: Then we can wait for ihar overnight, I have a +2 on it, and he could merge it. | 21:15 |
| emagana | marun: I have addressed all Ihar's comments | 21:15 |
| clarkb | fwiw I could find no documentation on what services are required when using dvr on a compute node so that was all trial and error and lots of log reading :) | 21:15 |
| *** luqas has quit IRC | 21:15 | |
| armax | clarkb: there is a reference architecture on the Wiki | 21:15 |
| armax | clarkb: but yes, we clearly forgot about fw and meta agents | 21:16 |
| clarkb | armax: have a link? I found some things on the wiki but they were very "bland" | 21:16 |
| clarkb | eg devoid of real info | 21:16 |
| armax | clarkb: https://wiki.openstack.org/wiki/Neutron/DVR | 21:16 |
| armax | that one missed the fw agent, as the fw agent is optional | 21:16 |
| armax | depending on whether you have firewall in the deployment or not | 21:17 |
| clarkb | armax: it doesn't cover metadata service | 21:17 |
| armax | it’s right there | 21:17 |
| armax | metadata | 21:17 |
| clarkb | and I think its still an open question if dhcp should run there too, but dhcp seems to work without that (I think because l2 broadcast on br-int works to the controller) | 21:17 |
| armax | underneath l3(dvr) on the compute node | 21:17 |
| clarkb | armax: gah its in a picture | 21:17 |
| armax | :) | 21:17 |
| clarkb | armax: honestly that should be greppable... | 21:18 |
| armax | well it’s a reference architecture diagram | 21:18 |
| *** tiswanso_ has quit IRC | 21:18 | |
| clarkb | I ^F meta and if its not there close the link | 21:18 |
| armax | that could’ve done ascii | 21:18 |
| clarkb | or just describe in words what the picture is trying to convey | 21:19 |
| clarkb | and have the picture along as helpful illustration | 21:19 |
| armax | a picture can tell more than a 100 words | 21:19 |
| armax | but I guess that’s not hte point | 21:19 |
| *** SumitNaiksatam has quit IRC | 21:21 | |
| emagana | Thanks Ihar! | 21:22 |
| emagana | just one more bug to close! | 21:22 |
| clarkb | armax: I have held the nodes running https://jenkins02.openstack.org/job/check-tempest-dsvm-neutron-aiopcpu-full/47/console so we can get on there and poke around when tempest is done | 21:22 |
| *** dkehn has joined #openstack-neutron | 21:22 | |
| *** dboik_ has quit IRC | 21:23 | |
| armax | clarkb: so do you think we should not try and spend any time figuring out whether neutron_dvr_gre_bridge from swami’s patch is viable? | 21:23 |
| armax | or at least to use it as a comparison term? | 21:23 |
| clarkb | armax: my change is basically that | 21:23 |
| *** busterswt has quit IRC | 21:23 | |
| clarkb | armax: but it fixes a couple bugs in it | 21:23 |
| clarkb | armax: it switches nova net to using the ovs gre bridge | 21:24 |
| armax | clarkb: ic, so are you going to later ‘deprecate’ gre_bridge? | 21:25 |
| *** Sayaji has joined #openstack-neutron | 21:25 | |
| clarkb | armax: if this works, yes I think that is the sanest thing to do | 21:25 |
| armax | clarkb: ok | 21:25 |
| clarkb | armax: I am not super tied to any implementation as long as neutron and nova-net can share one in d-g | 21:25 |
| *** dkehn has quit IRC | 21:25 | |
| armax | clarkb: 100% in agreement there | 21:26 |
| clarkb | we can even give vxlan a go, but I am still somewhat in the dark about what the multicast requirements are (which we can't rely on in cloud VMs) | 21:26 |
| *** shashankhegde has quit IRC | 21:26 | |
| armax | clarkb: as for multicast, Neutron does not rely on it | 21:27 |
| *** obondarev_ has joined #openstack-neutron | 21:27 | |
| HenryG | pc_m: In neutron it seems the test module is not discovered and therefore never run :( cc: marun | 21:27 |
| armax | clarkb: you mean for the tunnel setup? | 21:27 |
| *** bjornar_ has quit IRC | 21:27 | |
| clarkb | armax: yes | 21:27 |
| clarkb | armax: instead of using gre | 21:27 |
| *** obondarev has quit IRC | 21:27 | |
| armax | clarkb: if gre works well, and we have two nodes only I am not sure what’s the value in switching to vxlan | 21:28 |
| *** rushil_ has quit IRC | 21:28 | |
| pc_m | HenryG: That was my guess. I'm testing my change right now. Seems like only one case of the renamed ut. | 21:28 |
| *** rushil has joined #openstack-neutron | 21:29 | |
| *** s3wong has joined #openstack-neutron | 21:30 | |
| mestery | armax clarkb: Not to throw fire on things, but I'd pick vxlan if I was starting from scratch too :) | 21:30 |
| *** ronis has quit IRC | 21:30 | |
| clarkb | mestery: armax can anyone fill me in on when multicast is required? | 21:30 |
| mestery | clarkb: It's not required nor supported with OVS VXLAN. | 21:31 |
| mestery | clarkb: If you're using LB, it's optional | 21:31 |
| armax | clarkb: afaik never | 21:31 |
| armax | for ovs | 21:31 |
| mestery | clarkb: But honestly, multicast ... really? | 21:31 |
| mestery | :) | 21:31 |
| clarkb | mestery: so my reading has said multicast is how vxlan does broadcasts within logical l2 networks | 21:31 |
| *** jorgem has quit IRC | 21:31 | |
| clarkb | mestery: both dhcp and arp rely on ^ | 21:31 |
| clarkb | is that an implementation specific detail and not specific to vxlan itself? | 21:32 |
| mestery | clarkb: It coudl, but it's not required. | 21:32 |
| mestery | clarkb: We know the MAC to IP mappings, it's a constrained environment, so we pre-populate them | 21:32 |
| clarkb | ok so as a generic thing vxlan won't work? | 21:32 |
| clarkb | because we need to support nova network too and aiui its not going to prepopulate anything | 21:33 |
| mestery | clarkb: I'm not following how vxlan multicast is related to nova-network support here. | 21:33 |
| *** mwagner_lap has joined #openstack-neutron | 21:33 | |
| clarkb | mestery: because we have to set up a test env level logical layer 2 network for both nova net testing and neutron testing on >1 node | 21:34 |
| clarkb | mestery: for nova net this is a "flat" network and a "public" network that spans our test nodes | 21:34 |
| clarkb | mestery: for neutron this is the "br-ex" network | 21:34 |
| mestery | clarkb: Yes, for neutron all you need is L3 connectivity | 21:34 |
| mestery | they don't even have to be the same network | 21:34 |
| pc_m | emacs !$ | 21:35 |
| clarkb | but for nova network we must have l2 | 21:35 |
| pc_m | :) | 21:35 |
| clarkb | so we are doing neutron the same | 21:35 |
| marun | HenryG: which test module? | 21:35 |
| clarkb | because I am not going to support 2 different network topologies if I can avoid it (also I should be able to replace nova net on a cloud with neutron without rewiring my hardware) | 21:35 |
| *** rpothier has quit IRC | 21:36 | |
| clarkb | mestery: but in addition to that we don't have l3 control either | 21:36 |
| clarkb | mestery: so its just simpler with an l2 overlay all around | 21:36 |
| mestery | clarkb: It can work with what you've described | 21:36 |
| *** rushil has quit IRC | 21:36 | |
| pc_m | marun: neutron/tests/functional/agent/test_l3_agent.py | 21:37 |
| clarkb | mestery: well initial tests say it doesn't :) but the goal is to make it work for sure | 21:37 |
| pc_m | HenryG: marun: In my repo, that test module is run (and fails on the import). | 21:37 |
| clarkb | what appears to happen is the l3 agent adds the floating ip, arpings the floating ip with no response, removes the floating ip, doesn't tell nova about it, tempest then times out trying to ssh to the floating ip | 21:38 |
| marun | pc_m: | 21:39 |
| *** greghaynes has joined #openstack-neutron | 21:39 | |
| *** s3wong has quit IRC | 21:40 | |
| pc_m | marun: yes? | 21:40 |
| *** banix has quit IRC | 21:40 | |
| marun | pc_m: what's the error? | 21:40 |
| marun | pc_m: I'm working on reproducing locally | 21:40 |
| armax | clarkb: about that, I am still trying to find where the logic is | 21:40 |
| armax | I wonder if carl_baldwin knows a bit more, since he’s the resident l3-agent expert | 21:41 |
| pc_m | marun: There is an invalid import of unit test. | 21:41 |
| clarkb | armax: oh I have logs uploaded to the log server now which might give you a clearer picture of that. Let me get you a link | 21:41 |
| *** Sukhdev has quit IRC | 21:41 | |
| armax | carl_baldwin: are you aware of such a behavior? | 21:41 |
| marun | pc_m: ah, I guess the reorg must have missed it | 21:41 |
| marun | pc_m: :( | 21:41 |
| * carl_baldwin getting context. Reading back... | 21:41 | |
| clarkb | armax: http://logs.openstack.org/14/171714/5/experimental/check-tempest-dsvm-neutron-aiopcpu-full/a5d0c03/ that test run | 21:41 |
| clarkb | carl_baldwin: ^ | 21:41 |
| *** johnbelamaric has quit IRC | 21:41 | |
| marun | pc_m, HenryG: I'd like to know why the failure doesn't stop testing in its tracks. | 21:41 |
| *** dane_leblanc has quit IRC | 21:42 | |
| marun | lifeless: shouldn't import errors fail discovery? | 21:42 |
| clarkb | marun: yes they should | 21:42 |
| pc_m | marun: yeah, that's what is baffling me, why we didn't see it. I found it trying VPN code that imported the functional test module. | 21:42 |
| marun | clarkb: have you seen cases where failure was silent, as we're seeing/ | 21:43 |
| marun | ? | 21:43 |
| clarkb | marun: no thats new to me | 21:43 |
| marun | :( | 21:43 |
| kfox1111 | hmm... ok. one question related to option 42... what is a user to do when they want to force an emediate time set at boot? does ntpdate work with option 42 also? | 21:44 |
| *** claired has joined #openstack-neutron | 21:44 | |
| kfox1111 | what about right after unsuspending the vm? | 21:44 |
| kfox1111 | I've had to ansible ntpdate <ip> after those types of events. | 21:45 |
| haleyb | kfox1111: as a data point, how is that different from when not using dhcp for ntp-servers? | 21:46 |
| *** jckasper has joined #openstack-neutron | 21:47 | |
| kfox1111 | if the ntp server was always 169.254.169.254 the commands are a lot simpler. | 21:47 |
| kfox1111 | its a known command. | 21:47 |
| pc_m | marun: I think the check script we have only checks the unit test portion of the repo, IIRC | 21:50 |
| marun | pc_m: check script? | 21:50 |
| *** vivek-ebay has joined #openstack-neutron | 21:51 | |
| pc_m | marun: never mind, that only catches module naming... | 21:51 |
| kfox1111 | I'm not sure there is an ntpdate command you can issue that says, use whatever ntp server dhcp recommends. | 21:51 |
| clarkb | armax: the nova-net run with the ovs gre bridge just passed, so this works for nova net | 21:51 |
| pc_m | marun: not the imports | 21:52 |
| armax | clarkb: https://jenkins02.openstack.org/job/check-tempest-dsvm-neutron-aiopcpu-full/47/console | 21:52 |
| armax | looking here…and it seems it’s looking good so far | 21:52 |
| carl_baldwin | clarkb, sorry for the delay. Juggling too many things. There is a work-around in the L3 agent when running arping from a fip namespace. It will add the fip, arping, and then remove the fip. | 21:52 |
| *** claired has quit IRC | 21:52 | |
| *** jckasper has quit IRC | 21:52 | |
| clarkb | carl_baldwin: my assertion is that is a bug | 21:52 |
| clarkb | carl_baldwin: at the very least it needs to tell nova | 21:52 |
| marun | pc_m, HenryG: It's my fault for not checking, but the functional job should have failed | 21:52 |
| clarkb | carl_baldwin: so that the error is clear | 21:52 |
| clarkb | carl_baldwin: but it makes it almost impossible to debug neutron when it cleans up after itself like that | 21:53 |
| marun | pc_m, HenryG: I see the problem now... | 21:53 |
| marun | lifeless: hi | 21:53 |
| carl_baldwin | clarkb, the thing is that the fip ns doesn't really need the IP configured on its interface except to work around this arping bug. | 21:53 |
| clarkb | carl_baldwin: all I know is the fip doesn't work | 21:53 |
| pc_m | marun: I have a fix tested. Should I hold on pushing, in case we want to do another change? | 21:53 |
| clarkb | carl_baldwin: it is removed so no interface has it and no NAT happens | 21:53 |
| *** vivek-ebay has quit IRC | 21:53 | |
| pc_m | marun: for test run detecting the failure? | 21:53 |
| clarkb | carl_baldwin: which makes DVR broken from tempest perspective | 21:53 |
| *** vivek-ebay has joined #openstack-neutron | 21:53 | |
| marun | I have the reason for the discovery issue | 21:54 |
| clarkb | armax: https://jenkins07.openstack.org/job/check-tempest-dsvm-aiopcpu-full/121/console is the successful nova-net + ovs gre bridge run | 21:54 |
| carl_baldwin | clarkb, Does the IP appear in the qrouter namespace? | 21:54 |
| marun | pc_m, HenryG: https://github.com/openstack/neutron/blob/master/neutron/tests/functional/__init__.py | 21:54 |
| clarkb | carl_baldwin: I don't know yet, waiting for tmpest to finish running then I can poke around on the environment | 21:54 |
| *** amotoki has joined #openstack-neutron | 21:54 | |
| clarkb | carl_baldwin: https://jenkins02.openstack.org/job/check-tempest-dsvm-neutron-aiopcpu-full/47/console is the job | 21:54 |
| marun | pc_m: this needs to be removed | 21:54 |
| *** crose has quit IRC | 21:54 | |
| clarkb | armax: and ya its good so far, usually fails at the end when it starts ssh into nodes, but maybe the earlier failures were a fluke? | 21:55 |
| marun | pc_m: you mean you have a fix for all the failing imports in the functional job? | 21:55 |
| marun | pc_m: there is more than just one | 21:55 |
| clarkb | armax: in any case i can try to reproduce on that env when the test completes | 21:55 |
| pc_m | marun: I only found one case of a unit test import in the functional tree that was wrong. Didn't check fullstack or others. | 21:55 |
| marun | pc_m: I have a more comprehensive fix in the works. | 21:56 |
| *** atuvenie has quit IRC | 21:56 | |
| clarkb | armax: I am seeing a couple tests that previously failed pass in that run | 21:56 |
| clarkb | armax: so maybe it is working | 21:56 |
| pc_m | marun: By removed, do you mean the whole __init__.py | 21:56 |
| marun | pc_m: yeah, we can't rely on that mechanism | 21:56 |
| kfox1111 | yeah. ntpdate has to have a server specified... | 21:56 |
| lifeless | marun: hi | 21:56 |
| marun | lifeless: hi | 21:56 |
| pc_m | marun: OK. So should I abandon work on this. | 21:56 |
| marun | lifeless: We're seeing something strange around some manual test discovery we were doing: https://github.com/openstack/neutron/blob/master/neutron/tests/api/__init__.py | 21:57 |
| marun | pc_m: Thank you for discovering this. I'll have a patch that fixes my error and the discovery issues with functional and api. | 21:57 |
| marun | lifeless: using this method for discovery, invalid imports in a given test module resulted in that test module being silently ignored | 21:58 |
| clarkb | armax: if it does work we can run the tests a few times to see if it is consistent, and I can clean up the change to remove the old gre bridge stuff | 21:58 |
| pc_m | marun: OK | 21:58 |
| lifeless | do we still support 2.6? If so that should be unittest2 that you import | 21:58 |
| armax | clarkb: sounds like a plan | 21:58 |
| marun | lifeless: we don't support 2.6, no | 21:58 |
| marun | lifeless: should I try that then? | 21:59 |
| * marun goes to try it | 21:59 | |
| haleyb | kfox1111: i don't see this as a neutron-specific thing, there's more than just an ntpdate that might need to happen on migration | 21:59 |
| armax | clarkb: it looks like the scenario tests are passing | 21:59 |
| armax | clarkb: those are the ones that do the ssh for sure | 21:59 |
| clarkb | armax: awesome, those are the most problematic ones. Yup | 21:59 |
| *** jorgem has joined #openstack-neutron | 22:00 | |
| *** kbringard1 has quit IRC | 22:01 | |
| *** chlong has joined #openstack-neutron | 22:01 | |
| lifeless | marun: sec, still processing | 22:01 |
| lifeless | marun: your code is buggy | 22:02 |
| lifeless | marun: the top level dir needs to be the top of the namespace | 22:02 |
| marun | lifeless: ? | 22:03 |
| lifeless | marun: e.g. os.path.abspath(os.path.join(__file__, '../../../..')) or some such | 22:03 |
| lifeless | that said I don't understand why you're discarding the parameterised loader | 22:03 |
| marun | lifeless: so the fact that it's discovering tests that don't contain import errors is luck? | 22:03 |
| marun | lifeless: it's not my code, tbh | 22:03 |
| *** carl_baldwin has quit IRC | 22:03 | |
| lifeless | I think you need unittest2, because import handling was fixed in Python3.5 | 22:04 |
| clarkb | armax: shelve instance test just failed but that may be unrelated to the old failures | 22:04 |
| lifeless | marun: this is what I'd do: | 22:04 |
| clarkb | armax: I found that shelving instances is not the most reliable thing when getting it to work with nova net | 22:05 |
| *** seizadi has quit IRC | 22:05 | |
| lifeless | http://paste.openstack.org/show/200743/ | 22:06 |
| lifeless | or | 22:06 |
| marun | lifeless: ok, and the top_level_dir should be the parent dir of the neutron package? | 22:08 |
| *** annegentle has quit IRC | 22:09 | |
| lifeless | http://paste.openstack.org/show/200745/ | 22:10 |
| lifeless | ^ thats the entire file, no import of unittest etc. | 22:10 |
| lifeless | give that a go | 22:10 |
| marun | lifeless: will do, danke! | 22:10 |
| *** s3wong has joined #openstack-neutron | 22:10 | |
| lifeless | marun: who wrote the code? | 22:11 |
| *** jamielennox is now known as jamielennox|away | 22:11 | |
| lifeless | marun: also, why do you need this, since retargetable will be discovered *anyway* ? | 22:11 |
| marun | lifeless: jschwarz | 22:11 |
| pc_m | marun: so there is the same issue in VPN repo https://github.com/openstack/neutron-vpnaas/blob/master/neutron_vpnaas/tests/functional/openswan/__init__.py | 22:11 |
| marun | lifeless: he used this as a way of adding multiple test paths to a single tox env run | 22:12 |
| armax | clarkb: if we can at least demonstrate that some ssh tests do pass | 22:12 |
| lifeless | marun: but they're both under neutron/tests | 22:12 |
| *** shashankhegde has joined #openstack-neutron | 22:12 | |
| armax | clarkb: then we can safely say that the ttunnelling setup is sound | 22:12 |
| marun | lifeless: we isolate tests by path | 22:12 |
| openstackgerrit | John Perkins proposed openstack/neutron: Non-json body on POST 500's https://review.openstack.org/171836 | 22:12 |
| armax | and we can move on to the next troubleshooting phase | 22:12 |
| marun | lifeless: neutron/tests/unit requires minimal setup | 22:12 |
| marun | lifeless: neutron/tests/functional, retargetable and fullstack require devstack setup | 22:13 |
| marun | lifeless: neutron/tests/api requires devstack-deployed neutron | 22:13 |
| marun | lifeless: so the goal here was to allow logical separation via paths of different kinds of testing, but allow them to run as part of the same tox env | 22:14 |
| lifeless | marun: I don't understand the problem, just discover them all and filter down to what you need | 22:14 |
| marun | lifeless: filter them how? | 22:14 |
| lifeless | for instance: testr run neutron.tests.api neutron.tests.functional | 22:14 |
| armax | clarkb: which is, figuring out Neutron’s glitches :) | 22:15 |
| lifeless | will run both of those subdirs | 22:15 |
| marun | lifeless: the way our tox.ini was dealing with things may be the limiting factor that led us to this solution then. let me revisit. | 22:15 |
| clarkb | armax: yup, the thing to be careful of there is tempest runs on the controller node so we specifically want to prove that ssh from controller node to the sub node compute node works | 22:15 |
| armax | clarkb: that’s a good point | 22:16 |
| armax | clarkb: we should try to remove n-cpu from the controller? | 22:16 |
| clarkb | armax: no, I don't think so. We will just have to cross check nodes being booted on the subnode against the tests that ssh | 22:16 |
| openstackgerrit | Kevin Benton proposed openstack/neutron: Add simple ARP spoofing protection https://review.openstack.org/171003 | 22:16 |
| armax | clarkb: I meant as a WIP, not for good | 22:17 |
| clarkb | armax: oh maybe, we can see how hard it is to cross check the ssh tests against the compute node first (since that doesn't require rerunning the jobs with new code) | 22:17 |
| armax | ok | 22:17 |
| armax | kevinbenton: ping | 22:18 |
| clarkb | armax: ok the latest run is down to 4 failures (from ~13) so thats an improvement | 22:18 |
| armax | indeed | 22:18 |
| kevinbenton | armax: pong | 22:18 |
| clarkb | armax: but the failures look like ssh timeouts too. I am going to try manually adding a node on that setup and see if I can reproduce | 22:18 |
| armax | kevinbenton: can you answer my question on PS10 for patch https://review.openstack.org/#/c/171003/ | 22:18 |
| pc_m | marun: I have to log off, but can you add me to the review of the commit you post, so that I can see if I need to do a similar solution in VPN repo? (not sure about FW and LB) | 22:19 |
| armax | kevinbenton: not sure if you’ve seen it | 22:19 |
| kevinbenton | armax: oh, i missed it, one sec | 22:19 |
| marun | ok | 22:19 |
| *** banix has joined #openstack-neutron | 22:19 | |
| *** dttocs has joined #openstack-neutron | 22:19 | |
| marun | lifeless: so testr accepts multiple args... | 22:19 |
| *** annegentle has joined #openstack-neutron | 22:20 | |
| marun | lifeless: our tox uses env vars to configure .testr.conf in a way that doesn't appear to accept multiple args, though | 22:20 |
| marun | lifeless: https://github.com/openstack/neutron/blob/master/.testr.conf#L2 | 22:20 |
| *** mancdaz has quit IRC | 22:21 | |
| armax | clarkb: the l3 agent log for the subnode is awfully tiny | 22:22 |
| kevinbenton | armax: just left a comment | 22:22 |
| kevinbenton | armax: let me know if it makes sense | 22:22 |
| armax | kevinbenton: looking | 22:22 |
| *** mancdaz has joined #openstack-neutron | 22:22 | |
| *** kickinz1 is now known as kickinz1|afk | 22:22 | |
| armax | kevinbenton: it kinda does, but at the moment nothing use the map table besides the arp spoofing rule | 22:23 |
| armax | kevinbenton: so what’s the benefit of having it in? | 22:23 |
| marun | lifeless: supplying multiple arguments results in and error: 'Path must be within the project' | 22:23 |
| marun | lifeless: e.g. OS_TEST_PATH=./neutron/tests/functional ./neutron/tests/fullstack ./neutron/tests/retargetable | 22:23 |
| clarkb | armax: ok, I have manually botted a node, updated security-groups to allow icmp and tcp 22, and attached a floating ip and I cannot ping or ssh to that floating ip | 22:24 |
| *** moha_hunt has quit IRC | 22:24 | |
| openstackgerrit | John Perkins proposed openstack/neutron: Non-json body on POST 500's https://review.openstack.org/171836 | 22:24 |
| kevinbenton | armax: there isn't anything specific to the arp rules in the map table | 22:24 |
| lifeless | marun: whats OS_TEST_PATH | 22:24 |
| armax | clarkb: that is on teh subnode? | 22:24 |
| kevinbenton | armax: it's just a thing to keep track of ports -> ofports | 22:24 |
| kevinbenton | armax: if something changes, it retriggers the setup | 22:24 |
| lifeless | marun: oh, set OS_TEST_PATH to . | 22:24 |
| marun | lifeless: as per the link to our tox.ini, it's how our tox env's communicate the desired path to testr | 22:24 |
| clarkb | armax: yes according to virsh list --all the VM is on the subnode | 22:24 |
| lifeless | marun: you want to discover everything | 22:24 |
| marun | lifeless: ? | 22:25 |
| *** seizadi has joined #openstack-neutron | 22:25 | |
| kevinbenton | armax: right now arp spoofing is the only thing that benefits since it's the only 'in_port' based rules | 22:25 |
| marun | lifeless: we don't want to run unit tests | 22:25 |
| clarkb | armax: curiously I cannot ping the br-ex interface on the controller from the subnode either. But can ping the br-ex interface on the subode from the controller | 22:25 |
| kevinbenton | armax: but if we add more in the future, this would automatically help them as well | 22:25 |
| clarkb | hrm br-ex on subnode doesn't have an address anymore | 22:25 |
| clarkb | please don't tell me that neutron is cleaning that up... | 22:25 |
| lifeless | marun: right - so you'll select what you want specifically | 22:25 |
| lifeless | marun: discovered != run | 22:26 |
| marun | lifeless: https://github.com/openstack/neutron/blob/master/.testr.conf#L2 | 22:26 |
| marun | lifeless: I think we're already doing that | 22:26 |
| armax | clarkb: um I am afraid so | 22:26 |
| marun | lifeless: discover -t ./ ${OS_TEST_PATH...} | 22:26 |
| armax | clarkb: um..no hang on | 22:26 |
| armax | clarkb: I am getting confused | 22:26 |
| lifeless | marun: you're not, because you are setting OS_TEST_PATH to neutron/tests/thing | 22:27 |
| marun | lifeless: but ./ is the first entry? | 22:27 |
| lifeless | marun: -t ./ <- top dir | 22:27 |
| marun | lifeless: *sigh* | 22:27 |
| kevinbenton | armax: i'm on a train and might lose IRC connection. if i don't respond, just leave a gerrit comment | 22:27 |
| armax | kevinbenton: ok, it’s feels like two changes in one, but I understand your rationale | 22:27 |
| *** prad has quit IRC | 22:27 | |
| marun | lifeless: can you point me to a project that does this correctly? | 22:27 |
| kevinbenton | armax: well this does require the ofport protection | 22:27 |
| lifeless | marun: testr.conf is not intended to select tests to run | 22:27 |
| *** ivar-lazzaro has quit IRC | 22:27 | |
| kevinbenton | armax: i just made it generic to help other future stuff | 22:27 |
| lifeless | marun: AFAIK every other project | 22:27 |
| marun | lifeless: ok, looking | 22:28 |
| *** achanda has quit IRC | 22:28 | |
| armax | kevinbenton: that I understand, I don’t undestand why we can remove the check when we indeed need this piece of code | 22:28 |
| marun | lifeless: nova does the same thing as we do | 22:28 |
| marun | lifeless: we probably just followed them | 22:28 |
| lifeless | it does? ugh | 22:28 |
| *** aepifanov has quit IRC | 22:28 | |
| armax | kevinbenton: what scares me is that part of the code is not at all tested | 22:28 |
| lifeless | clarkb: where did this pattern come from ? | 22:28 |
| marun | lifeless: so does keystone :/ | 22:28 |
| lifeless | marun: so - OS_TEST_PATH shouldn't exist. | 22:28 |
| marun | lifeless: ok. | 22:29 |
| armax | and having it run pointlessly now begs the question | 22:29 |
| lifeless | marun: it should just be . (or perhaps PROJECTNAME) | 22:29 |
| clarkb | lifeless: what pattern? the test path thing? thats because they wanted to colocate tests | 22:29 |
| *** seizadi has quit IRC | 22:29 | |
| lifeless | clarkb: but discovery != selection, we can filter what to run - tox -epy27 -- neutron.tests.api | 22:29 |
| lifeless | for instance | 22:29 |
| kevinbenton | armax: hmmm. let me see if i can get a test for it | 22:30 |
| clarkb | yes I had no part in this | 22:30 |
| clarkb | lifeless: don't look at me | 22:30 |
| lifeless | clarkb: yeah, but you see everything happen | 22:30 |
| *** shashankhegde has quit IRC | 22:30 | |
| armax | kevinbenton: it’ll be a lot easier and more digestable if you made it conditional | 22:30 |
| lifeless | clarkb: so I'm wondering who can tell me what was unclear about the docs / architecture that lead them to this :) | 22:30 |
| armax | kevinbenton: and remove the condition, when indeed this becomes useful for other purposes | 22:30 |
| marun | lifeless: I think everyone does it this way :( | 22:30 |
| morganfainberg | it might also have been an artifact of when tests were ./tests and moved to ./{PROJECT}/tests | 22:30 |
| lifeless | morganfainberg: defintely not | 22:30 |
| openstackgerrit | enikanorov proposed openstack/neutron: Add logging to dangling port to ml2 delete_subnet https://review.openstack.org/171662 | 22:30 |
| morganfainberg | lifeless, no? | 22:30 |
| lifeless | morganfainberg: *that* clarkb and infra did most of | 22:31 |
| kevinbenton | armax: that won't make it any more tested though :) | 22:31 |
| *** achanda has joined #openstack-neutron | 22:31 | |
| *** zhenq has quit IRC | 22:31 | |
| lifeless | marun: ok, so my suggestion is: | 22:31 |
| morganfainberg | lifeless, hm. | 22:31 |
| marun | lifeless: ok, so OS_TEST_PATH goes away | 22:31 |
| kevinbenton | armax: but if you want to ruin my code, i can definitely do that ;) | 22:31 |
| lifeless | 1) OS_TEST_PATH goes away | 22:31 |
| clarkb | morganfainberg: no that was done so that tests would get packaged and be available for use elsewhere iirc | 22:31 |
| armax | kevinbenton: true, but we’re really pushing the boundaries of what is and isn’t an RC1 | 22:31 |
| armax | bug | 22:31 |
| lifeless | 2) the custom discover stuff disappears (it has to, otherwise you'll get duplicate id errors) | 22:31 |
| morganfainberg | clarkb, yeah i remember when we did it in keystone... was ugly | 22:31 |
| lifeless | 3) to run different things you use one or more test patterns | 22:31 |
| clarkb | armax: let me know when you have time for me to data dump for you | 22:31 |
| marun | morganfainberg: I'm happy we did it, though. I've worked with projects that don't and it makes development env more complex. | 22:32 |
| marun | morganfainberg: can only run things from project root in some projects, which makes editor-invoked debugging sessions more complicated. | 22:32 |
| *** dttocs has quit IRC | 22:32 | |
| marun | lifeless: so how to pass test patterns? | 22:32 |
| armax | clarkb: flood me | 22:32 |
| *** banix has quit IRC | 22:32 | |
| lifeless | marun: I'm happy to help, though I need to wrap up some WIP before doing code myself - but it should be pretty straight forward | 22:32 |
| lifeless | marun: tox -- pattern1 pattern2 etc | 22:33 |
| lifeless | marun: see the {posargs} in tox.ini | 22:33 |
| marun | lifeless: we need it to be part of the env definition, though | 22:33 |
| marun | e.g. 'tox -e py27' without patterns | 22:33 |
| lifeless | marun: sure, so you put in in the same way that posargs is being passed | 22:33 |
| marun | lifeless: ah, ok | 22:33 |
| *** Sayaji has quit IRC | 22:33 | |
| clarkb | armax: ok, I was slightly confused myself earlier. So one of the things we do is put 172.24.4.1/23 on controller nodes br-ex and 172.24.4.2/23 on compute node's br-ex. Then we make the floating range 172.24.5.0/24. This automagically sets up networking such that controller can talk to nodes on compute node and vice versa. Problem is that neutron seems to actively subvert that | 22:34 |
| clarkb | armax: so I think part of the current issue is when neutron does ^ and breaks routing | 22:34 |
| clarkb | armax: so I am not surprised that I cannot ssh to nodes on the compute host from the controller, there is no route there so it will use my default route out into rackspaces network space | 22:35 |
| armax | clarkb: I wonder if the br-ex setup is not actively done on the subnode because devstack does not recognize the presence of the controller on the subnode? | 22:35 |
| clarkb | armax: well we are doing the br-ex setup because devstack doesn't do it properly on its own | 22:36 |
| clarkb | armax: but to continue on | 22:36 |
| amotoki | kevinbenton: ping | 22:36 |
| clarkb | armax: neutron does seem to put 172.24.5.1/24 on br-ex of the controller | 22:36 |
| *** annegentle has quit IRC | 22:36 | |
| clarkb | armax: so packets can make it from controller to compute host but not necessarily back again | 22:37 |
| kevinbenton | amotoki: pong | 22:37 |
| kevinbenton | amotoki: (i'm on an unreliable connection so i might be slow to respond) | 22:37 |
| lifeless | marun: and/or we can also teach discover to discover multiple directories | 22:37 |
| amotoki | kevinbenton: I noticed recent BigSwitch CI run no tempest tests. | 22:37 |
| lifeless | marun: and avoid all the glue | 22:37 |
| openstackgerrit | enikanorov proposed openstack/neutron: Handle race condition on subnet-delete https://review.openstack.org/171848 | 22:37 |
| kevinbenton | amotoki: oh, i'll check it out. thanks for letting me know | 22:38 |
| amotoki | kevinbenton: for example http://openstack-ci-gw.bigswitch.com/logs/refs-changes-17-167017-9/BSN_PLUGIN/testr_results.htmlgz | 22:38 |
| enikanorov | kevinbenton: armax: could you please provide some feedback on https://review.openstack.org/171848 ? | 22:38 |
| clarkb | armax: which makes me think there may be a bug somewhere (config, q-l3, etc) that isn't setting up an appropriate IP on br-ex of the compute node. Maybe thats what you meant about devstack setup being incomplete? | 22:38 |
| marun | lifeless: I'm guessing that's a slightly longer-term option | 22:38 |
| armax | clarkb: yes | 22:38 |
| armax | clarkb: that’s what I am trying to look for | 22:39 |
| clarkb | armax: gotcha | 22:39 |
| *** baoli_ has quit IRC | 22:39 | |
| lifeless | marun: few days, if we choose to do it - needs one patch to cPython, backport to unittest2, done. | 22:39 |
| *** alexpilotti has quit IRC | 22:39 | |
| lifeless | marun: I think what I proposed above is better though | 22:39 |
| *** baoli has joined #openstack-neutron | 22:39 | |
| marun | lifeless: I'm happy to do as you suggest, though it looks like it will require us duplicating our pretty tox invocation for every job :/ | 22:40 |
| marun | lifeless: https://github.com/openstack/neutron/blob/master/tox.ini#L18 | 22:40 |
| openstackgerrit | Merged openstack/neutron: Re-use context session in ML2 DB get_port_binding_host https://review.openstack.org/170690 | 22:40 |
| armax | clarkb: fishing through the setup files in devstack to pinpoint where we do the br-ex interface setup | 22:41 |
| clarkb | armax: thanks, let me know if thee is anything I can do to help | 22:41 |
| marun | lifeless: or maybe I can reuse OS_TEST_PATH to parametize the invocation within tox? | 22:41 |
| lifeless | marun: nah, just use a setenv for the filter and put that in the actual invocation line | 22:41 |
| marun | lifeless: right, ok | 22:41 |
| lifeless | marun: needs a new variable, because its an object path not a file path | 22:41 |
| marun | lifeless: ok | 22:42 |
| *** thedodd has quit IRC | 22:42 | |
| clarkb | armax: and not to overload you with datapoints I see the arp who-has 172.24.5.243 (the floating ip) come into the compute node from the controller but no response even on the compute node | 22:42 |
| openstackgerrit | Merged openstack/neutron: Add missing config parameters in neutron.conf https://review.openstack.org/171059 | 22:42 |
| armax | clarkb: look here: | 22:43 |
| armax | http://logs.openstack.org/14/171714/6/experimental/check-tempest-dsvm-neutron-aiopcpu-full/b6b7e04/logs/devstacklog.txt.gz#_2015-04-08_21_27_57_253 | 22:43 |
| armax | clarkb: this isn’t done on the subnode | 22:43 |
| armax | clarkb: http://logs.openstack.org/14/171714/6/experimental/check-tempest-dsvm-neutron-aiopcpu-full/b6b7e04/logs/10.176.128.180-subnode/devstacklog.txt.gz | 22:43 |
| clarkb | armax: interesting. Looks like it happens as part of the neutron network creation which is all on the controller | 22:44 |
| armax | clarkb: yes, let me poke at this a bit more | 22:44 |
| clarkb | armax: so that is likely a devstack bug, however it should respect my IPs of 172.24.4.1/23 and 172.24.4.2/23 on the same interfaces right? | 22:44 |
| * clarkb double checks on ^ | 22:44 | |
| armax | clarkb: I don’t think that Neutron clears that | 22:44 |
| clarkb | armax: it might be devstack | 22:44 |
| armax | clarkb: if that’s what you’re saying | 22:45 |
| clarkb | ya its definitely cleared out by something, not sure what yet | 22:45 |
| armax | clarkb: right | 22:45 |
| clarkb | armax: http://logs.openstack.org/14/171714/6/experimental/check-tempest-dsvm-neutron-aiopcpu-full/b6b7e04/console.html#_2015-04-08_21_15_57_450 and http://logs.openstack.org/14/171714/6/experimental/check-tempest-dsvm-neutron-aiopcpu-full/b6b7e04/console.html#_2015-04-08_21_15_46_494 is where we add those IPs, the adds look fine | 22:46 |
| armax | clarkb: I think what happens is this | 22:46 |
| armax | clarkb: you mean added to the subnode right? | 22:47 |
| *** salv-orlando has quit IRC | 22:47 | |
| *** annegentle has joined #openstack-neutron | 22:47 | |
| clarkb | armax: we add 172.24.4.1/23 on the controller and 172.24.4.2/23 on the subnode | 22:47 |
| armax | clarkb: I think what happens is that devstack destroy and recreates br-ex | 22:47 |
| armax | that’s why the ip disappears | 22:47 |
| *** salv-orlando has joined #openstack-neutron | 22:48 | |
| armax | clarkb: as for the controller | 22:48 |
| armax | clarkb: the ip gets put back after we create the public network that supports the floating ip range | 22:48 |
| *** Marga_ has quit IRC | 22:48 | |
| clarkb | so we probably want to update devstack to do that for the subnode as well | 22:48 |
| clarkb | then we can avoid setting any addrs in the /23 space | 22:48 |
| armax | clarkb: look for _neutron_get_ext_gw_interface in lib/neutron-legacy | 22:49 |
| armax | clarkb: but on the subnode we don’t run that command at all | 22:49 |
| armax | clarkb: cuz the network is already created for us | 22:49 |
| clarkb | armax: yup | 22:49 |
| marun | lifeless: this doesn't seem to work: sh tools/pretty_tox.sh '{env:OS_TEST_FILTERS}' '{posargs}' | 22:50 |
| *** Marga_ has joined #openstack-neutron | 22:50 | |
| *** salv-orlando has quit IRC | 22:50 | |
| armax | clarkb: we definitely clear the br-ex prior to doing anything | 22:50 |
| armax | see neutron_ovs_base_cleanup in lib/neutron-legacy | 22:50 |
| armax | clarkb: hence the IP disappears | 22:51 |
| clarkb | armax: but that doesn't seem to clear the gre tunnel between the two br-exs | 22:51 |
| clarkb | armax: maybe that config isn't cleared as part of the set up? | 22:51 |
| clarkb | armax: probably the next step here is for me to clean up my changes and push to swami's as a new patchset | 22:51 |
| marun | lifeless: I'm setting OS_TEST_FILTERS via setenv, but it errors out saying that OS_TEST_FILTERS cannot be found :/ | 22:51 |
| clarkb | armax: then talk to devstack about how they want ot handle the setup | 22:51 |
| armax | clarkb: probably not | 22:51 |
| clarkb | armax: since I think we are going to need better devstack support | 22:51 |
| armax | clarkb: moving your changes to swami’s patch and just have a change that flips the DVR flag would be useful at this point | 22:52 |
| armax | clarkb: I think we have quite a bit of clarity that we’re getting on the right track | 22:52 |
| clarkb | armax: yup | 22:52 |
| armax | clarkb: let me see if I can pinpoint the culrpit in devstack itself | 22:52 |
| armax | clarkb: you certainly know your way around this stuff! | 22:53 |
| armax | clarkb: kudos | 22:53 |
| *** baoli has quit IRC | 22:53 | |
| clarkb | armax: well I spent quite a bit of time doing battle with nova net to get it working with multihost | 22:53 |
| clarkb | armax: many of the problems are mirrored here, it just goes through different code paths (both in devstack and neutron) | 22:53 |
| clarkb | armax: the nice thing about this is no ebtables \o/ | 22:53 |
| clarkb | (I think) | 22:54 |
| armax | clarkb: not yet :P | 22:54 |
| *** jorgem has quit IRC | 22:54 | |
| *** sbfox has joined #openstack-neutron | 22:54 | |
| marun | lifeless: I could do the evil thing and pick up OS_TEST_FILTERS in pretty tox | 22:55 |
| *** mancdaz has quit IRC | 22:57 | |
| armax | clarkb: yup we definitely flush the bridge br-ex | 22:58 |
| clarkb | armax: thanks for confirming | 22:58 |
| armax | clarkb: right before we setup the l3-agent | 22:58 |
| armax | in _neutron_ovs_base_configure_l3_agent | 22:58 |
| armax | we could in theory add a flag there | 22:59 |
| armax | like if MULTI_HOST=0 | 22:59 |
| armax | then follow the usual setup | 22:59 |
| armax | if not, then shut up and move on | 22:59 |
| clarkb | armax: no, I think its better if devstack just does the setting of the interface IP for all nodes running routers | 22:59 |
| armax | so to speak> | 22:59 |
| armax | oh, that’d be even better | 22:59 |
| clarkb | armax: then I can avoid setting 172.24.4.1/23 and 172.24.4.2/23 at all | 22:59 |
| armax | right! | 22:59 |
| clarkb | armax: and it should work more better for individuals deploying this locally | 22:59 |
| armax | the 172.24.4.1/23 is done on the controller already | 23:00 |
| clarkb | armax: no thats 172.24.5.1/24 | 23:00 |
| armax | clarkb: right, got it | 23:00 |
| *** pc_m has quit IRC | 23:00 | |
| clarkb | with nova net the trick we used was to have an encapsulating network that the controller and compute node say on in order to have routing set up properly for us | 23:00 |
| *** absubram has joined #openstack-neutron | 23:00 | |
| clarkb | otherwise packets bound for the VM floating ips would go out the default route which was eth0 and onto rax/hpcloud networks | 23:01 |
| armax | clarkb: right 172.24.4.1/23 == tunnel network | 23:01 |
| clarkb | yup | 23:01 |
| armax | clarkb: got confused there for a sec | 23:01 |
| armax | I am with you | 23:01 |
| clarkb | but in this case we just need an l2 tunnel network then devstack/neutron can manage the l3 atop that | 23:01 |
| clarkb | devstack to configure interfaces on hypervisors and neutron to configure floating ips | 23:02 |
| armax | clarkb: ok so to be clear you really wanna get rid of command on line 888 for https://review.openstack.org/#/c/171714/6/functions.sh | 23:03 |
| *** annegentle has quit IRC | 23:03 | |
| armax | as well as the one on line 874? | 23:03 |
| armax | clarkb: and want these done in devstack itself? | 23:03 |
| clarkb | armax: ya but only for neutron I think | 23:03 |
| clarkb | armax: I have a patch almost together, give me a couple minutes and I can show you | 23:04 |
| kevinbenton | armax: can't get a functional test that fast for that | 23:04 |
| armax | clarkb: for devstack? | 23:04 |
| kevinbenton | armax: so i will make it conditional and just rely on UT for now | 23:04 |
| clarkb | armax: for devstack-gate | 23:04 |
| armax | clarkb: ok | 23:04 |
| *** annegentle has joined #openstack-neutron | 23:05 | |
| *** chlong has quit IRC | 23:05 | |
| clarkb | armax: https://review.openstack.org/171714 | 23:05 |
| marun | lifeless: I don't really understand why it's a good idea to discover all tests if we know at the outset that we only want to run a subset. | 23:05 |
| armax | kevinbenton: I offered you a much simpler way and yet you want to do it the hard way, even when it’s wrong | 23:05 |
| armax | clarkb: looking | 23:05 |
| marun | lifeless: but I do have your suggestion working. | 23:06 |
| openstackgerrit | Kevin Benton proposed openstack/neutron: Add simple ARP spoofing protection https://review.openstack.org/171003 | 23:06 |
| clarkb | armax: but for that change to work properly we also need a devstack change to set the interfaces for us | 23:06 |
| kevinbenton | armax: what? | 23:07 |
| armax | clarkb: ok, but just to be clear, why do we set them in devstack-gate too? | 23:07 |
| kevinbenton | armax: didn't i do exactly what you wanted | 23:07 |
| kevinbenton | armax: ? | 23:08 |
| armax | kevinbenton: I suggested to make it conditional as it’s useless right now | 23:08 |
| *** mlavalle has quit IRC | 23:08 | |
| armax | kevinbenton: with or without the test | 23:08 |
| clarkb | armax: we set them during nova net bceause devstack does not set them when using nova net, and when you use nova net there is no reserved space for routers of the floating ips | 23:08 |
| armax | clarkb: ok | 23:08 |
| kevinbenton | armax: it's not useless when the spoofing rules are enabled | 23:08 |
| clarkb | armax: so for nova net floating ip routing we put 172.24.4.1/23 and 172.24.4.2/23 on the public interfaces so that packets know how to go back and forth for IPs on that network | 23:09 |
| armax | kevinbenton: but it’s disabled right now is it not? | 23:09 |
| kevinbenton | armax: which is why i was trying to write a test | 23:09 |
| clarkb | armax: then we restrict floating IPs to 172.24.5.0/24 | 23:09 |
| armax | kevinbenton: ok fine | 23:09 |
| clarkb | armax: but neutron appears to have slightly different expectations and devstack is already half setting it up for us, so might as well have it do the whole thing | 23:09 |
| *** sbfox has quit IRC | 23:09 | |
| armax | clarkb: so you’re expecting to set another ip space for br-ex | 23:10 |
| clarkb | armax: also the reason non DVR works today on multinode is that we only connect from controller to compute host VMs but all routing for compute host VMs happens on the controller | 23:10 |
| armax | clarkb: but done through devstack | 23:10 |
| clarkb | armax: ya something like that | 23:10 |
| clarkb | armax: or if devstack prefers we can do it in d-g then make devstack not interfere with the settings we do in d-g | 23:10 |
| clarkb | but its interfering so it can do it :) | 23:10 |
| armax | clarkb: either way yes, we need a change in devstack | 23:11 |
| clarkb | yup | 23:11 |
| armax | clarkb: but it sounds like that we’d still need the setup in d-g regardless | 23:11 |
| clarkb | armax: we still need to do the l2 tunnel setup regardless | 23:11 |
| armax | clarkb: right | 23:11 |
| clarkb | armax: whether or not we assign 172.24.4.1/23 and 172.24.4.2/23 is largely up to devstack | 23:11 |
| armax | clarkb: let me push a devsatck patch to see how this might look like and we can take the conversastion there | 23:12 |
| armax | clarkb: indeed | 23:12 |
| *** jamielennox|away is now known as jamielennox | 23:12 | |
| clarkb | armax: sounds good | 23:12 |
| *** luqas has joined #openstack-neutron | 23:12 | |
| *** gildub has joined #openstack-neutron | 23:13 | |
| *** mancdaz has joined #openstack-neutron | 23:14 | |
| *** annegentle has quit IRC | 23:16 | |
| *** dims has quit IRC | 23:16 | |
| *** luqas has quit IRC | 23:16 | |
| armax | clarkb: just a question, but the variables set in d-g like FLOATING_HOST_MASK etc | 23:17 |
| armax | clarkb: are they visible to the env under which stash.sh works? | 23:17 |
| openstackgerrit | Assaf Muller proposed openstack/neutron: Log RPC initialization in L3 service plugin and ML2 https://review.openstack.org/171870 | 23:17 |
| clarkb | armax: thats purefly for the *_gre_bridge functions in functions.sh. They are not visible to stack.sh | 23:17 |
| amuller | marun: ^ | 23:17 |
| clarkb | *purely | 23:17 |
| clarkb | armax: thats another thing we might be able to make a better interface for | 23:18 |
| armax | clarkb: ok | 23:18 |
| clarkb | armax: I mostly hacked that together beause doing bit math in bash is no fun | 23:18 |
| openstackgerrit | Assaf Muller proposed openstack/neutron: Log RPC initialization in L3 service plugin and ML2 https://review.openstack.org/171870 | 23:20 |
| *** ZZelle_ has quit IRC | 23:20 | |
| *** absubram has quit IRC | 23:21 | |
| *** vivek-ebay has quit IRC | 23:22 | |
| lifeless | marun: its keeps the layering separate | 23:22 |
| openstackgerrit | Maru Newby proposed openstack/neutron: Fix test discovery https://review.openstack.org/171874 | 23:22 |
| marun | lifeless: ^^ | 23:22 |
| lifeless | marun: the internal glue that was being used to bang separate dirs together was non-obvious | 23:23 |
| marun | lifeless: so long as it works :) | 23:24 |
| marun | lifeless: I couldn't get the {env:} support to work, so I resorted to modifying pretty-tox.sh | 23:24 |
| *** emagana has quit IRC | 23:25 | |
| *** srampal has quit IRC | 23:25 | |
| *** hichihara has joined #openstack-neutron | 23:28 | |
| lifeless | marun: I've mailed openstack-dev vis-a-vis not using 'unittest'. | 23:29 |
| amuller | marun: replied on the patch | 23:29 |
| lifeless | I don't know if its worth a hacking rule, but it might be | 23:29 |
| *** amotoki has quit IRC | 23:29 | |
| *** yamamoto has joined #openstack-neutron | 23:30 | |
| marun | lifeless: danke | 23:31 |
| *** yamamoto_ has joined #openstack-neutron | 23:31 | |
| *** yamamoto has quit IRC | 23:31 | |
| *** sbfox has joined #openstack-neutron | 23:32 | |
| *** baoli has joined #openstack-neutron | 23:32 | |
| *** dims has joined #openstack-neutron | 23:34 | |
| openstackgerrit | Maru Newby proposed openstack/neutron: Fix test discovery https://review.openstack.org/171874 | 23:34 |
| clarkb | armax: I left a long comment on https://review.openstack.org/#/c/158525/ to catch swami up to what we figured out today | 23:34 |
| *** amittal has joined #openstack-neutron | 23:35 | |
| clarkb | lifeless: marun if you just depend on testtools and use testtools.TestCase unittest2 will sort itself out right? | 23:36 |
| lifeless | clarkb: context? | 23:36 |
| *** seizadi has joined #openstack-neutron | 23:36 | |
| marun | lifeless: I think the problem is using unittest for discovery, not for test cases | 23:36 |
| clarkb | lifeless: the don't use unittest thing | 23:36 |
| marun | oops, clarkb ^ | 23:36 |
| lifeless | clarkb: the particular issue here was discovery | 23:37 |
| clarkb | but discovery will just use whatever you import | 23:37 |
| lifeless | clarkb: but we shouldn't be directly using unittest anywhere | 23:37 |
| clarkb | so import testtools | 23:37 |
| lifeless | clarkb: no, discovery is also an API | 23:37 |
| lifeless | clarkb: see the diff in the review | 23:37 |
| *** seizadi has quit IRC | 23:37 | |
| lifeless | clarkb: api/__init__.py | 23:37 |
| *** seizadi1 has joined #openstack-neutron | 23:37 | |
| *** dboik_ has joined #openstack-neutron | 23:37 | |
| lifeless | clarkb: yes, we should be using testtools, and testtools works with both unittest and unittest2 and so on | 23:38 |
| clarkb | oh I didn't realize it was doing manual test loads | 23:38 |
| *** yeungp has quit IRC | 23:38 | |
| lifeless | clarkb: *if* we're using a unittest api directly, for whatever reason, we should use the one from unittest2. | 23:38 |
| clarkb | gotcha | 23:39 |
| lifeless | because otherwise we'll be getting random old bugs that I've already fixed | 23:39 |
| *** seizadi1 has quit IRC | 23:41 | |
| *** amuller has quit IRC | 23:42 | |
| *** sbfox has quit IRC | 23:43 | |
| *** apuimedo has quit IRC | 23:44 | |
| *** Sukhdev has joined #openstack-neutron | 23:44 | |
| Sukhdev | armax: Ping | 23:44 |
| armax | Sukhdev: pong | 23:45 |
| Sukhdev | armax: wanted to check with you if you are OK with moving L2-GW meeting one hour early to accomodate India team? | 23:46 |
| *** sbfox has joined #openstack-neutron | 23:46 | |
| Sukhdev | armax: Did not see your reply to that email - hence, thought I ping you here | 23:46 |
| armax | Sukhdev: I shoudl be ok | 23:53 |
| Sukhdev | armax: OK - let check which channel is available - will send out email with information | 23:53 |
| *** kfox1111 has quit IRC | 23:54 | |
| Sukhdev | armax: Also, I wanted to document the versioning and release plan for the L2-GW project - what is the right place for it? | 23:54 |
| *** psd has quit IRC | 23:54 | |
| *** amittal has quit IRC | 23:54 | |
| *** watanabe_isao has joined #openstack-neutron | 23:55 | |
| *** rha has quit IRC | 23:58 | |
| *** nplanel has quit IRC | 23:58 | |
| *** rha has joined #openstack-neutron | 23:59 | |
| *** rha has joined #openstack-neutron | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
